Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Amide Proton Transfer-weighted MRI combined with serum prostate-specific antigen levels for differentiating malignant prostate lesions from benign prostate lesions: a retrospective cohort study

Amide Proton Transfer-weighted MRI combined with serum prostate-specific antigen levels for... Background Early diagnosis of prostate cancer improves its prognosis, while it is essential to upgrade screening tools. This study aimed to explore the value of a novel functional magnetic resonance imaging (MRI) technique, namely amide proton transfer (APT )-weighted MRI, combined with serum prostate-specific antigen (PSA) levels to differentiate malignant prostate lesions from benign prostate lesions. Methods Data of patients who underwent prostate examinations at Chongqing University Cancer Hospital between July 2019 and March 2022 were retrospectively analyzed. All patients underwent T2-weighted imaging ( T2WI), APT, diffusion-weighted imaging (DWI), and dynamic contrast-enhanced (DCE) MRI. Two radiologists analyzed the images independently. The ability of the quantitative parameters alone or in different combinations in differentiating malig- nant prostate lesions from benign prostate lesions were compared by using receiver operating characteristic (ROC) curves. According to the DeLong test, the combined parameters were significantly different from the corresponding single parameter (P < 0.05). Results A total of 79 patients were finally enrolled, including 52 patients in the malignant group and 27 patients in the benign group. The separate assessment of indexes revealed that APTmax, APTmean, mean apparent diffusion coefficient (ADCmean), ADCmax, ADCmin, tPAD, free prostate-specific antigen (FPSA), FPSA/total prostate-specific antigen (tPSA), and PSA density (PSAD) were significantly different between the two groups (P < 0.05), while APTmin was not significantly different between the two groups (P > 0.05). APTmax and APTmean had the high values of area under the ROC curve (AUC), which were 0.780 and 0.710, respectively. APTmax had a high sensitivity, and APTmean had a high specificity. The combination of APTmax, APTmean, ADCmean, and PSAD had the highest AUC value (AUC: 0.880, sensitivity: 86.540, specificity: 78.260). Lu Yang and Lei Wang contributed equally to this study. *Correspondence: Jiuquan Zhang zhangjq_radiol@foxmail.com Full list of author information is available at the end of the article © The Author(s) 2023. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/. The Creative Commons Public Domain Dedication waiver (http:// creat iveco mmons. org/ publi cdoma in/ zero/1. 0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data. Yang et al. Cancer Imaging (2023) 23:3 Page 2 of 8 Conclusion APTmax, APTmean, ADCmean, ADCmin, tPAD, FPSA, and PSAD showed to have a high value in differenti- ating malignant prostate lesions from benign prostate lesions in the separate assessment of indexes. The combination of APTmax, APTmean, ADCmean, and PSAD had the highest diagnostic value. Keywords Amide proton transfer (APT )-weighted MRI, Prostate-specific antigen, Prostate cancer, Malignant lesions, Benign lesions Introduction parameters, which could be used as imaging biomarkers Prostate cancer is the second most common cancer in for diverse types of cancer, thereby facilitating the early men worldwide and the most common cancer in men diagnosis of malignant lesions. Serum prostate-specific in the United States [1, 2]. The annual age-standardized antigen (PSA) is the biochemical biomarker that has incidence of prostate cancer is 29.3 per 100,000 men been highly acknowledged in clinical practice for prostate for an estimated 1,276,106 cases and 358,989 deaths in diseases [13]. Nevertheless, this biomarker is not per- 2018 [3]. The incidence of prostate cancer has gradually fect [14, 15], and its combination with other biomarkers increased in the recent decades [4]. The median age at the could result in better sensitivity and specificity for pros - time of diagnosis of prostate cancer is 66  years old. The tate cancer. overall five-year survival rate of patients with prostate It is noteworthy that APT-weighted MRI has advanced cancer is noticeable 98% [5]. However, the five-year sur - rapidly in recent years [16–19]. As an endogenous chemi- vival rate of localized prostate cancer and prostate cancer cal exchange saturation transfer (CEST) imaging tech- with extraglandular and distant metastasis is remarkably nique and a molecular MRI imaging technique, APT different. The five-year survival rate for prostate cancer imaging is considered as the most clinically feasible CEST with extraglandular and distant metastasis is 31%. There - imaging because its specific resonance frequency is dif - fore, the Gleason score and pathological T stage are two ferent from the water resonance frequency, and it allows of the most important prognostic factors for prostate the exchange of a large number of water molecules with cancer. Early diagnosis and treatment could improve the amide protonic peptides in endogenous mobile proteins quality of life and five-year survival rate of patients with to acquire images [16, 17, 19]. APT imaging is based on prostate cancer [6]. the fact that tumor cells have a high proliferation activ- At present, multiparametric magnetic resonance imag- ity and a high protein synthesis, leading to differences in ing (mpMRI) is considered as one of the most effective protein amounts between benign and malignant lesions imaging methods in diagnosing prostate cancer. It is [19]. Therefore, acquiring the quantitative APT param - noteworthy that MRI has a high soft-tissue resolution eters could quantitatively reflect the molecular changes and allows the determination of the Prostate Imaging and could be used as a potential imaging biomarker. Reporting & Data System (PI-RADS) score [7], differ - Therefore, the present study aimed to explore the value entiation of malignant lesions from benign lesions, pre- of the APT-weighted MRI combined with serum PSA operative evaluation of malignant lesions, and evaluation levels to differentiate malignant prostate lesions from of treatment efficacy [8, 9]. Nevertheless, the sensitivity benign prostate lesions, and to provide more imaging and specificity of MRI for prostate cancer can still be information for early preoperative diagnosis of prostate improved [10, 11]. A recent study explored the diagnostic cancer. value of combination of amide proton transfer (APT) and mpMRI in transition zone (TZ) prostate cancer. Differ - Methods ences in APT-weighted and apparent diffusion coefficient Study design and patients (ADC) values between TZ prostate cancer and benign Data of patients who underwent prostate examinations prostatic hyperplasia (BPH), and differences in T2* values at Chongqing University Cancer Hospital (Chongqing, between stromal BPH and glandular BPH were found. It China) between July 2019 and March 2022 were ret- was found that APT-weighted and ADC were independ- rospectively analyzed. This study was approved by the ent predictors of TZ prostate cancer. Moreover, a combi- Ethics Committee of Chongqing University Cancer Hos- nation of APT and ADC values improved the diagnostic pital. The requirement for patients’ informed consent was sensitivity of TZ prostate cancer and achieved the pur- waived by the Ethics Committee. pose of improving the diagnostic efficiency [12]. All patients underwent T2-weighted imaging Novel MRI techniques are limited to conventional (T2WI), APT, diffusion-weighted imaging (DWI), and medical imaging, present the details of morphological dynamic contrast-enhanced (DCE) MRI. The inclu- features of lesions, and innovatively provide functional sion criteria were as follows: 1) untreated patients with Y ang et al. Cancer Imaging (2023) 23:3 Page 3 of 8 prostate diseases, and 2) patients who received MRI independent-samples t-test or the Wilcoxon rank-sum examination. The exclusion criteria were as follows: test. Two-sided P < 0.05 was considered statistically sig- 1) history of prostate surgery or endocrine therapy, nificant. Receiver operating characteristic (ROC) curves or 2) patients who underwent prostate biopsy within were plotted for the analysis. For parameters with statisti- 4–6 weeks before MRI. cally significant differences, the logistic regression analy - sis was used to estimate the probability of the combined MRI examination parameters. The parameters with statistically significant An INGENIA 3.0  T MRI scanner (Philips Healthcare differences were modeled using the binary logistic regres - Co., Ltd., Best, The Netherlands) was used for the MRI sion analysis, and the probability of the combined param- of prostate of all patients from July 2019 to March eters was calculated. ROC curves were plotted to compare 2022, including APT, DWI, T2WI, and DCE, using a the combinations of parameters in differentiating malig - 32-channel phased-array body coil (Table 1). The scan- nant prostate lesions from benign prostate lesions. ning parameters included b = 0 and 1400  s/mm for The index with the highest diagnostic accuracy was DWI and SPIR for fat suppression in APT. selected from the separate assessment of APT/ADC and PSA indexes to establish the combined model. The Image analysis SPSS software was employed to select the binary logis- The images were processed by the ISP post-processing tic regression model, and two or more parameters were workstation (Philips Healthcare Co., Ltd.). Two radi- obtained through modeling, in order to jointly predicting ologists reviewed the images independently and blindly. the probability of malignant tumors. The maximum lesion level was selected using the T2-weighted images as the standard. Then the region of Results interest (ROI) was selected at the same level on the APT General characteristics of patients and DWI images to measure the APT and ADC values A total of 123 patients were recruited, of whom 44 (Fig.  1). The averages of values measured by the two patients were excluded according to the exclusion cri- radiologists were calculated for the statistical analysis. teria. Finally, 79 patients were included in this study, of Patients’ medical records were reviewed to obtain the which 52 patients were in the malignant group (patholog- total PSA (tPSA) and free PSA (FPSA). ically proven with prostate cancer [20]) and 27 patients were in the benign group (pathologically proven with Statistical analysis non-cancerous lesions, including prostatic hyperplasia SPSS 23.0 software (IBM, Armonk, NY, US) was used and prostatitis). The mean age was 70.54 ± 8.66 years old. to perform the statistical analysis. Intraclass correlation There were no significant differences in age and volume (ICC) analysis was performed for the measurement data of prostate between malignant group and benign group obtained by the two radiologists (ICC > 0.75 indicated (P < 0.05) (Table 2). a high consistency, ICC equal to 0.4–0.75 indicated a moderate consistency, and ICC < 0.4 indicated a low con- Consistency between radiologists sistency). The Kolmogorov–Smirnov test was used to ICC consistency was performed for all the measurement assess the normal distribution, and the Levene’s test was data by the two radiologists. As shown in Table  3, APT employed for the homogeneity assessment of the con- and ADC parameters showed a moderate consistency tinuous data. Continuous data were analyzed using the between the two radiologists. Table 1 Parameters of the scanning sequences Sequences TR (ms) TE (ms) TA Layers Layer Interlayer FOV (mm) Matrix NSA (s) thickness spacing (mm) (mm) T2WI 3000 110 24 3 0.3 200 × 200 308 × 255 1 TSE factor 14 1′56’’ DWI 4828 90 3′8’’ 24 3 0 200 × 200 80 × 81 1 EPI factor 55 APT 5842 7.9 4′54’’ 18 5 0 140 × 140 80 × 78 1 TSE factor 174 T1WI 4.0 2.0 6′24’’ 24 3.5 0 250 × 250 180 × 140 1 dynamic phase 80 TR repetition time, TE echo time, TA acquisition time, FOV field of view, NSA number of signals per acquisition, T2WI T2-weighted imaging, DWI diffusion-weighted imaging, APT amide proton transfer, T1WI T1-weighted imaging TSE turbo spin echo, EPI echo-planar imaging Yang et al. Cancer Imaging (2023) 23:3 Page 4 of 8 Fig. 1 Case examples. 1 A 75-year-old man was hospitalized, and the physical examination revealed that PSA level increased. (1A) T2WI. (1B) APT. (1C) DWI. (1D) ADC. 2 A 68-year-old man was hospitalized for treatment after finding the possibility of prostate malignancy in the external hospital. (2A) T2WI. (2B) APT. (2C) DWI. (2D) ADC Ability of the quantitative parameters in differentiating malignant prostate lesions from benign prostate lesions Table 3 ICC consistency test of the results measured by the two radiologists As presented in Table  4, the quantitative parameters in differentiating malignant prostate lesions from benign Radiologist 1 Radiologist 2 ICC 95% CI prostate lesions were significantly different (P < 0.05), APTmean (%) 2.44 ± 1.21 2.42 ± 1.38 0.732 0.649–0.789 except for APTmin (P > 0.05). Quantitative APT param- APTmax (%) 5.56 ± 2.12 6.06 ± 2.26 0.701 0.605–0.768 eters, including APTmax and APTmean, could well dis- APTmin (%) -1.61 ± 2.33 -1.55 ± 2.62 0.652 0.538–0.734 tinguish malignant prostate lesions from benign prostate –3 ADCmean (10 0.91 ± 0.32 0.88 ± 0.37 0.664 0.554–0.743 lesions. Comparatively, APTmax had a higher sensitivity mm /s) (92.310), and APTmean had a higher specificity (77.780), –3 ADCmax (10 1.55 ± 0.45 1.47 ± 0.51 0.629 0.507–0.718 while they both had high area under the curve (AUC) val- 2 mm /s) ues (Fig.  1A and Table  5).The other parameters had rela - –3 ADCmin (10 0.53 ± 0.34 0.61 ± 0.53 0.566 0.395–0.699 tively high values in differentiating malignant prostate mm /s) lesions from benign prostate lesions. ROC curves were ICC intraclass correlation, CI confidential interval, APT amide proton transfer, ADC apparent diffusion coefficient Table 2 Characteristics of the patients Characteristics (n = 79) Malignant (n = 52) Benign (n = 27) P value Age (years) 69.29 ± 8.20 72.19 ± 9.43 0.42 Volume of prostate (cm ) 9.34 ± 12.46 10.83 ± 14.66 0.22 PSAD (ng/mL/cm ) 69.73 ± 137.59 3.76 ± 7.83 0.00 Gleason score, n (%) Low-risk subgroup (≤ 3 + 3 points) 15 (28.85%) High-risk subgroup (> 3 + 3 points) 37 (71.15%) PSAD prostate-specific antigen density Y ang et al. Cancer Imaging (2023) 23:3 Page 5 of 8 Table 4 Ability of the quantitative parameters in differentiating malignant from benign prostate diseases Malignant (n = 52) Benign (n = 27) Z/t P value APTmean (%) 2.740 ± 1.106 1.859 ± 1.200 -3.263 0.002 APTmax (%) 6.698 ± 1.837 4.530 ± 1.894 -4.924 < 0.001 APTmin (%) -1.788 ± 2.557 -1.259 ± 1.815 0.965 0.342 –3 2 ADCmean (10 mm /s) 0.845 (0.300–2.030) 1.190 (0.620–1.680) -4.259 < 0.001 –3 2 ADCmax (10 mm /s) 1.450 ± 0.445 1.738 ± 0.391 2.835 0.006 –3 2 ADCmin (10 mm /s) 0.464 ± 0.356 0.646 ± 0.271 2.330 0.022 tPSA (ng/mL) 33.604 (0.600–8169.000) 6.053 (0.840–131.760) -3.321 0.001 FPSA (ng/mL) 3.959 (0.020–483.000) 1.230 (0.210–7.240) -2.506 0.012 FPSA/tPSA 16.159 ± 15.300 23.336 ± 12.954 2.074 0.043 PSAD (ng/mL/cm ) 10.585 (0.010–688.540) 1.980 (0.030–38.650) -2.884 0.004 APT amide proton transfer, ADC apparent diffusion coefficient, tPSA total prostate-specific antigen, FPSA free prostate-specific antigen, PSAD prostate-specific antigen density Table 5 Analysis of the ROC curves of the different quantitative parameters AUC 95%CI Sensitivity Specificity Youden index J Best cut-off value P value APTmax 0.780 0.673–0.865 92.310 59.260 0.516 4.300 < 0.001 APTmean 0.710 0.598–0.807 63.460 77.780 0.412 2.300 0.001 ADCmean 0.793 0.688–0.876 75.000 81.480 0.565 1.060 < 0.001 ADCmax 0.689 0.575–0.788 46.150 92.580 0.388 1.360 0.002 ADCmin 0.689 0.575–0.789 73.080 70.370 0.435 0.620 0.003 tPSA 0.734 0.628–0.836 57.690 91.300 0.490 28.140 < 0.001 FPSA 0.684 0.565–0.788 42.000 100.000 0.420 7.240 0.003 FPSA/tPSA 0.684 0.565–0.788 68.000 69.570 0.376 17.050 0.007 PSAD 0.700 0.594–0.809 53.850 95.650 0.495 8.210 < 0.001 ROC receiver operating characteristic, AUC area under the curve, CI confidence interval, APT amide proton transfer, ADC apparent diffusion coefficient, tPSA total prostate-specific antigen, FPSA free prostate-specific antigen, PSAD prostate-specific antigen density DeLong et al., [21] plotted and showed that ADCmean had the highest AUC the combined parameters were significantly different value (0.793, 95% confidence interval (CI): 0.688–0.876), from the corresponding single parameter (P < 0.05). APTmax had the highest sensitivity (92.310), and FPSA had the highest specificity (100.000) (Fig.  2A, B, C and Discussion Table 5). Early diagnosis of prostate cancer improves its prognosis [1, 2], while screening tools should be further upgraded Combinations of parameters in differentiating malignant [10, 11, 14, 15]. Therefore, the present study aimed to prostate lesions from benign prostate lesions explore the value of the novel functional MRI technique, As reported previously, APTmax, APTmean, and APT-weighted MRI, combined with serum PSA levels for ADCmean [22], the most representative parameters of differentiating malignant prostate lesions from benign APT and ADC, were used in combination with PSAD, prostate lesions. The results indicated that APTmax, which is the most representative parameter of PSA APTmean, ADCmean, ADCmax, ADCmin, tPAD, FPSA, [23]. The combinations of these quantitative param- FPSA/tPSA, and PSAD had a high clinical value in differ - eters in differentiating malignant prostate lesions from entiating malignant prostate lesions from benign prostate benign prostate lesions were compared, and the ROC lesions. The combination of APTmax, APTmean, ADC - curves were plotted. As illustrated in Fig.  2D, E, F and mean, and PSAD showed the highest diagnostic value. Table 6, the AUC value was the highest when APTmax, Early diagnosis has important clinical significance for APTmean, and ADCmean were combined with PSAD the treatment and prognosis of patients with prostate (AUC: 0.880, 95% CI: 0.784–0.943, sensitivity: 86.540, cancer. Nevertheless, the early differentiation of malig - specificity: 78.260). According to the DeLong test [21], nant prostate lesions from benign prostate lesions is Yang et al. Cancer Imaging (2023) 23:3 Page 6 of 8 Fig. 2 Receiver operating characteristic (ROC) curves. A ROC curves of APT parameters. B ROC curves of the DWI parameters. C ROC curves of the PSA parameters. D ROC curves of the combinations of APTmax, ADCmean, and PSAD. D ROC curves of the combinations of APTmean, ADCmean, and PSAD. E ROC curves of the combinations of APTmax, APTmean, ADCmean, and PSAD. APT: amide proton transfer; ADC: apparent diffusion coefficient; tPSA: total prostate-specific antigen; FPSA: free prostate-specific antigen; ratio: FPSA/tPSA. PSAD: prostate-specific antigen density Table 6 Analysis of the ROC curves of the combinations of different quantitative parameters AUC 95%CI Sensitivity Specificity Youden index J P value APTmax + ADCmean 0.841 0.763–0.928 92.310 70.370 0.627 < 0.001 APTmax + PSAD 0.826 0.721–0.904 80.770 69.570 0.503 < 0.001 ADCmean + PSAD 0.854 0.753–0.925 75.000 91.300 0.663 < 0.001 APTmean + ADCmean 0.837 0.748–0.918 78.850 85.190 0.640 < 0.001 APTmean + PSAD 0.826 0.721–0.904 75.000 82.960 0.576 < 0.001 APTmean + ADCmean + PSAD 0.878 0.782–0.942 80.770 91.300 0.721 < 0.001 APTmax + APTmean + ADCmean + PSAD 0.880 0.784–0.943 86.540 78.260 0.648 < 0.001 ROC receiver operating characteristic, AUC area under the curve, CI confidence interval, APT amide proton transfer, ADC apparent diffusion coefficient, tPSA total prostate-specific antigen, FPSA free prostate-specific antigen, PSAD prostate-specific antigen density DeLong et al., [21] still difficult in clinical practice based only on the cur - signal values, depending on the exchange ratio rent imaging methods, such as ultrasound and MRI. between the amide protons and free-water protons. The APT imaging technique is based on transferring The exchange ratio depends on pH values and protein the amide protons and water, and it reflects the changes concentrations in the body. The APT technique was of proteins and pH values by variations of water sig- initially used for the nervous system [24–26]. In recent nals. The internal contrast is acquired by measuring the years, a great number of researchers have applied the water signals to indirectly acquire the APT-weighted APT technique to diagnose prostate diseases [27, 28]. Y ang et al. Cancer Imaging (2023) 23:3 Page 7 of 8 The findings of the present study showed that APTmax respectively) [29]. Guo et al. [12] found that APTmean and had a high diagnostic value for differentiating malignant ADC were independent predictors of TZ prostate cancer. prostate lesions from benign prostate lesions. The sen - Moreover, combination of APTmean and ADC values sitivity was the highest, indicating that the maximum improved the sensitivity of the diagnosis of TZ prostate transfer of amide protons and exchange ratio of water cancer and achieved the purpose of improving the diag- protons in a lesion could sensitively reflect the occur - nostic efficiency, which are similar to the results of the pre - rence of malignant lesions. Jia et al. [27], for the first time, sent study. attempted to apply the APT imaging in prostate diseases, The advantage of APT-weighted MRI is that it is a 3D and reported the value of this technique in differentiating imaging technique. Compared with the conventional malignant prostate lesions from benign prostate lesions. two-dimensional (2D) APT technique [30], this tech- Takayama et  al. [28] found that the APT-weighted val- nique could scan multiple layers in a short time, acquire ues in prostate cancer patients with a Gleason score of the APT image of the whole prostate region, and provide 7 points were significantly higher than those of patients more comprehensive functional information. with other scores. These findings were generally in agree - There were several limitations in the present study. ment with our results. The metabolism in malignant First, the sample size was small and imbalance, and addi- prostate lesions is more active than in benign lesions. The tional studies are required to determine the exact diag- exchange ratio of protons is higher, which is consistent nostic value of APT for prostate cancer. Second, this was with the pathological features of malignant lesions. The a single-center study, and local practice biases could influ - differences of APTmax and APTmean, two parameters ence the results. Multicenter studies can not only increase acquired by APT imaging, were statistically significant. the sample size, but also mitigate the risk of bias. Last but In contrast, the difference of APTmin was not statistically not least, because of the small sample size, no direct com- significant, which could be associated with the fact that parison was performed among imaging techniques. APTmin expresses the lowest value of protein content in the region of interest, and the difference is not enough Conclusions to distinguish between benign and malignant lesions. APT has a diagnostic value for prostate cancer. APTmax, Importantly, the minimum exchange ratio of protons APTmean, ADCmean, ADCmin, tPAD, FPSA, and PSAD could not reflect the degree of metabolic activity. showed to have a high diagnostic value in differentiating The results of the present study revealed that the differ - malignant prostate lesions from benign prostate lesions. ences of APTmax, APTmean, ADCmean, ADCmax, ADC- The combination of APTmax, APTmean, ADCmean, and min, tPAD, FPSA, FPSA/tPSA, and PSAD were statistically PSAD had the highest diagnostic value. significant (P < 0.05), suggesting that these parameters had high diagnostic values in differentiating malignant prostate Abbreviations lesions from benign prostate lesions. Among these param- ADC Apparent diffusion coefficient eters, ADCmean had the highest AUC, APTmax had the APT Amide proton transfer AUC Ar ea under the curve highest sensitivity, and FPSA had the highest specificity CEST Chemical exchange saturation transfer and the lowest sensitivity, reflecting that FPSA had the DWI Diffusion-weighted imaging highest diagnostic accuracy and a relatively high false-neg- FPSA Free prostate-specific antigen ICC Intraclass correlation ative rate. This indicated that functional MRI sequences, MRI Magnetic resonance imaging such as DWI and PSA, can be used as independent pre- PI-RADS P rostate Imaging Reporting & Data System dictive biomarkers to discriminate benign prostate lesions PSA Prostate-specific antigen ROC Receiver operating characteristic and malignant prostate lesions. In contrast, APTmax had ROI Region of interest the highest positive rate. After combining these param- DCE T1-weighted dynamic contrast-enhanced (DCE) eters, the results showed that the combination of APTmax, T2WI T2-weighted imaging tPSA Total prostate-specific antigen APTmean, ADCmean, and PSAD had the highest AUC and sensitivity, suggesting that this combination had the Acknowledgements highest diagnostic value for prostate cancer. Compared The authors would like to thank all participants who were enrolled in this study. We highly appreciate the efforts dedicated by Dr. Xiaoyong Zhang from with using PSA and DWI alone, the combination of APT, Philips Healthcare (Clinical Science) for his assistance in APT-MRI sequence DWI, and PSA techniques had a relatively high diagnostic development and valuable discussions. value (AUC: 0.880) and a high sensitivity (86.540) for pros- Authors’ contributions tate cancer. A previous study demonstrated that the AUC Conceptualization: Jiuquan Zhang; Data curation: Yuchuan Tan; Formal values for the biparametric MRI (bpMRI) and multipara- analysis: Hanli Dan; Funding acquisition: Lu Yang; Investigation: Yipeng Zhang; metric MRI (mpMRI) protocols for prostate cancer were Methodology: Peng Xian; Project administration: Lu Yang; Resources: Meng Lin; Software: Yong Tan; Supervision: Jiuquan Zhang; Validation: Lu Yang; Roles/ comparable (0.790 [0.732–0.840] and 0.791 [0.733–0.841], Yang et al. Cancer Imaging (2023) 23:3 Page 8 of 8 Writing—original draft: Lei Wang; Writing—review and editing: Lu Yang. The 10. Zhen L, Liu X, Yegang C, Yongjiao Y, Yawei X, Jiaqi K, et al. Accuracy of mul- author(s) read and approved the final manuscript. tiparametric magnetic resonance imaging for diagnosing prostate Can- cer: a systematic review and meta-analysis. BMC Cancer. 2019;19(1):1244. Funding 11. Becker AS, Kirchner J, Sartoretti T, Ghafoor S, Woo S, Suh CH, et al. Interac- This study was supported by the National Natural Science Foundation of tive, up-to-date meta-analysis of MRI in the management of men with China (Grant No. 82071883), the General Project of Chongqing Natural Science suspected prostate cancer. J Digit Imaging. 2020;33(3):586–94. Foundation (Grant No. cstc2021jcyj-msxmX0447), the China National Cancer 12. Guo Z, Qin X, Mu R, Lv J, Meng Z, Zheng W, et al. Amide proton transfer Center Climbing Fund (Grant No. NCC201822872), and the Open Research could provide more accurate lesion characterization in the transition Fund Project of Chongqing University Cancer Hospital (2020). The funder had zone of the prostate. J Magn Reson Imaging. 2022;56(5):1311–9. the following roles in the study: study design, collection, analysis, and interpre- 13. Pezaro C, Woo HH, Davis ID. Prostate cancer: measuring PSA. Intern Med J. tation of data, writing the paper and decision to submit it for publication. 2014;44(5):433–40. 14. Ilic D, Djulbegovic M, Jung JH, Hwang EC, Zhou Q, Cleves A, et al. Prostate Availability of data and materials cancer screening with prostate-specific antigen (PSA) test: a systematic All data generated or analyzed during this study are included in this published review and meta-analysis. BMJ. 2018;362:k3519. article. 15. Lee SK, Kim J-Y, Jeong HS. Benign peripheral nerve sheath tumor of digit versus major-nerve: comparison of MRI findings. PLoS ONE. 2020;15(3):e0230816. Declarations 16. Meng N, Wang X, Sun J, Han D, Ma X, Wang K, et al. Application of the amide proton transfer-weighted imaging and diffusion kurtosis imaging Ethics approval and consent to participate in the study of cervical cancer. Eur Radiol. 2020;30(10):5758–67. This study was approved by the Ethics Committee of Chongqing Univer- 17. Kamitani T, Sagiyama K, Togao O, Yamasaki Y, Hida T, Matsuura Y, et al. sity Cancer Hospital, and it was performed in accordance with the ethical Amide proton transfer (APT ) imaging of parotid tumors: Differentiation of standards laid down in an appropriate version of the Declaration of Helsinki malignant and benign tumors. Eur J Radiol. 2020;129:109047. (as revised in Brazil 2013). The requirement for patients’ informed consent was 18. Ma X, Bai Y, Lin Y, Hong X, Liu T, Ma L, et al. Amide proton transfer mag- waived by the Ethics Committee. netic resonance imaging in detecting intracranial hemorrhage at differ - ent stages: a comparative study with susceptibility weighted imaging. Sci Consent for publication Rep. 2017;7:45696. Not applicable. 19. Sartoretti T, Sartoretti E, Wyss M, Schwenk A, Najafi A, Binkert C, et al. Amide proton transfer contrast distribution in different brain regions in Competing interests young healthy subjects. Front Neurosci. 2019;13:520. The authors declare that they have no competing interests. 20. Humphrey PA, Moch H, Cubilla AL, Ulbright TM, Reuter VE. The 21. DeLong, E, DeLong, D, Clarke-Pearson, D. Comparing the Areas under Author details Two or More Correlated Receiver Operating Characteristic Curves: A Department of Radiology, Chongqing University Cancer Hospital & Nonparametric Approach BIOMETRICS. 1988;44(3):837. Chongqing Cancer Institute & Chongqing Cancer Hospital, No.181 Hanyu 22. Saito S, Koyama Y, Ueda J, Hashido T. Relationship between apparent Road, Shapingba District, Chongqing 400030, China. Department of Urol- diffusion coefficient distribution and cancer grade in prostate cancer and ogy, Chongqing University Cancer Hospital & Chongqing Cancer Institute & benign prostatic hyperplasia. Diagnostics (Basel). 2022;12(2):525. Chongqing Cancer Hospital, Chongqing 400030, China. 23. Milkovic B, Dzamic Z, Pejcic T, Kajmakovic B, Nikolic D, Cirovic D, et al. Evaluation of free-to-total prostate specific antigen (F/T PSA), prostate Received: 21 March 2022 Accepted: 19 December 2022 specific antigen density (PSAD) and (F/T )/PSAD sensitivity on reduction of unnecessary prostate biopsies for patients with PSA in gray zone. Ann Ital Chir. 2014;85(5):448–53. 24. Kang XW, Xi YB, Liu TT, Wang N, Zhu YQ, Wang XR, et al. Grading of Glioma: combined diagnostic value of amide proton transfer weighted, References arterial spin labeling and diffusion weighted magnetic resonance imag- 1. Mottet N, Corford P, van den Bergh RCN, Briers E, De Santis M, Fanti S, ing. BMC Med Imaging. 2020;20(1):50. et al. European Association of Urology (EAU). Guidelines on prostate 25. Zhang Z, Zhang C, Yao J, Chen X, Gao F, Jiang S, et al. Protein-based cancer. Arnhem: EAU Guidelines Office; 2020. amide proton transfer-weighted MR imaging of amnestic mild cognitive 2. NCCN Clinical Practice Guidelines in Oncology (NCCN Guidelines). Pros- impairment. Neuroimage Clin. 2020;25:102153. tate Cancer. Version 3.2020. Fort Washington: National Comprehensive 26. Zhou J, Heo HY, Knutsson L, van Zijl PCM, Jiang S. APT-weighted MRI: Cancer Network; 2020. Techniques, current neuro applications, and challenging issues. J Magn 3. Bray F, Ferlay J, Soerjomataram I, Siegel RL, Torre LA, Jemal A. Global Reson Imaging. 2019;50(2):347–64. cancer statistics 2018: GLOBOCAN estimates of incidence and mor- 27. Jia G, Abaza R, Williams JD, Zynger DL, Zhou J, Shah ZK, et al. Amide pro- tality worldwide for 36 cancers in 185 countries. CA Cancer J Clin. ton transfer MR imaging of prostate cancer: a preliminary study. J Magn 2018;68(6):394–424. Reson Imaging. 2011;33(3):647–54. 4. Jemal A, Bray F, Center MM, Ferlay J, Ward E, Forman D. Global cancer 28. Takayama Y, Nishie A, Sugimoto M, Togao O, Asayama Y, Ishigami K, et al. statistics. CA Cancer J Clin. 2011;61(2):69–90. Amide proton transfer (APT ) magnetic resonance imaging of prostate 5. Siegel RL, Miller KD, Jemal A. Cancer statistics, 2020. CA Cancer J Clin. cancer: comparison with Gleason scores. MAGMA. 2016;29(4):671–9. 2020;70(1):7–30. 29. Xu L, Zhang G, Shi B, Liu Y, Zou T, Yan W, et al. Comparison of biparametric 6. Nyame YA, Gore JL. What goes up must come down: identifying truth and multiparametric MRI in the diagnosis of prostate cancer. Cancer from global prostate cancer epidemiology. Eur Urol. 2020;77(1):53–4. Imaging. 2019;19(1):90. 7. PI-RADS. Prostate Imaging - Reporting and Data System Version 21. 30. Nishie A, Takayama Y, Asayama Y, Ishigami K, Ushijima Y, Okamoto D, et al. Philadelphia: American College of Radiology; 2019. Amide proton transfer imaging can predict tumor grade in rectal cancer. 8. Winkel DJ, Wetterauer C, Matthias MO, Lou B, Shi B, Kamen A, et al. Magn Reson Imaging. 2018;51:96–103. Autonomous detection and classification of PI-RADS lesions in an MRI screening population incorporating multicenter-labeled deep learn- Publisher’s Note ing and biparametric imaging: proof of concept. Diagnostics (Basel). Springer Nature remains neutral with regard to jurisdictional claims in pub- 2020;10(11):951. lished maps and institutional affiliations. 9. Qi Z, Li W, Tan J, Wang C, Lin H, Zhou B, et al. Eec ff t of ginsenoside Rh(2) on renal apoptosis in cisplatin-induced nephrotoxicity in vivo. Phy- tomedicine. 2019;61:152862. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Cancer Imaging Springer Journals

Amide Proton Transfer-weighted MRI combined with serum prostate-specific antigen levels for differentiating malignant prostate lesions from benign prostate lesions: a retrospective cohort study

Download PDF
8 pages

Loading next page...
 
/lp/springer-journals/amide-proton-transfer-weighted-mri-combined-with-serum-prostate-f2DNQhsXbm

References (34)

Publisher
Springer Journals
Copyright
Copyright © The Author(s) 2023
eISSN
1470-7330
DOI
10.1186/s40644-022-00515-w
Publisher site
See Article on Publisher Site

Abstract

Background Early diagnosis of prostate cancer improves its prognosis, while it is essential to upgrade screening tools. This study aimed to explore the value of a novel functional magnetic resonance imaging (MRI) technique, namely amide proton transfer (APT )-weighted MRI, combined with serum prostate-specific antigen (PSA) levels to differentiate malignant prostate lesions from benign prostate lesions. Methods Data of patients who underwent prostate examinations at Chongqing University Cancer Hospital between July 2019 and March 2022 were retrospectively analyzed. All patients underwent T2-weighted imaging ( T2WI), APT, diffusion-weighted imaging (DWI), and dynamic contrast-enhanced (DCE) MRI. Two radiologists analyzed the images independently. The ability of the quantitative parameters alone or in different combinations in differentiating malig- nant prostate lesions from benign prostate lesions were compared by using receiver operating characteristic (ROC) curves. According to the DeLong test, the combined parameters were significantly different from the corresponding single parameter (P < 0.05). Results A total of 79 patients were finally enrolled, including 52 patients in the malignant group and 27 patients in the benign group. The separate assessment of indexes revealed that APTmax, APTmean, mean apparent diffusion coefficient (ADCmean), ADCmax, ADCmin, tPAD, free prostate-specific antigen (FPSA), FPSA/total prostate-specific antigen (tPSA), and PSA density (PSAD) were significantly different between the two groups (P < 0.05), while APTmin was not significantly different between the two groups (P > 0.05). APTmax and APTmean had the high values of area under the ROC curve (AUC), which were 0.780 and 0.710, respectively. APTmax had a high sensitivity, and APTmean had a high specificity. The combination of APTmax, APTmean, ADCmean, and PSAD had the highest AUC value (AUC: 0.880, sensitivity: 86.540, specificity: 78.260). Lu Yang and Lei Wang contributed equally to this study. *Correspondence: Jiuquan Zhang zhangjq_radiol@foxmail.com Full list of author information is available at the end of the article © The Author(s) 2023. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/. The Creative Commons Public Domain Dedication waiver (http:// creat iveco mmons. org/ publi cdoma in/ zero/1. 0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data. Yang et al. Cancer Imaging (2023) 23:3 Page 2 of 8 Conclusion APTmax, APTmean, ADCmean, ADCmin, tPAD, FPSA, and PSAD showed to have a high value in differenti- ating malignant prostate lesions from benign prostate lesions in the separate assessment of indexes. The combination of APTmax, APTmean, ADCmean, and PSAD had the highest diagnostic value. Keywords Amide proton transfer (APT )-weighted MRI, Prostate-specific antigen, Prostate cancer, Malignant lesions, Benign lesions Introduction parameters, which could be used as imaging biomarkers Prostate cancer is the second most common cancer in for diverse types of cancer, thereby facilitating the early men worldwide and the most common cancer in men diagnosis of malignant lesions. Serum prostate-specific in the United States [1, 2]. The annual age-standardized antigen (PSA) is the biochemical biomarker that has incidence of prostate cancer is 29.3 per 100,000 men been highly acknowledged in clinical practice for prostate for an estimated 1,276,106 cases and 358,989 deaths in diseases [13]. Nevertheless, this biomarker is not per- 2018 [3]. The incidence of prostate cancer has gradually fect [14, 15], and its combination with other biomarkers increased in the recent decades [4]. The median age at the could result in better sensitivity and specificity for pros - time of diagnosis of prostate cancer is 66  years old. The tate cancer. overall five-year survival rate of patients with prostate It is noteworthy that APT-weighted MRI has advanced cancer is noticeable 98% [5]. However, the five-year sur - rapidly in recent years [16–19]. As an endogenous chemi- vival rate of localized prostate cancer and prostate cancer cal exchange saturation transfer (CEST) imaging tech- with extraglandular and distant metastasis is remarkably nique and a molecular MRI imaging technique, APT different. The five-year survival rate for prostate cancer imaging is considered as the most clinically feasible CEST with extraglandular and distant metastasis is 31%. There - imaging because its specific resonance frequency is dif - fore, the Gleason score and pathological T stage are two ferent from the water resonance frequency, and it allows of the most important prognostic factors for prostate the exchange of a large number of water molecules with cancer. Early diagnosis and treatment could improve the amide protonic peptides in endogenous mobile proteins quality of life and five-year survival rate of patients with to acquire images [16, 17, 19]. APT imaging is based on prostate cancer [6]. the fact that tumor cells have a high proliferation activ- At present, multiparametric magnetic resonance imag- ity and a high protein synthesis, leading to differences in ing (mpMRI) is considered as one of the most effective protein amounts between benign and malignant lesions imaging methods in diagnosing prostate cancer. It is [19]. Therefore, acquiring the quantitative APT param - noteworthy that MRI has a high soft-tissue resolution eters could quantitatively reflect the molecular changes and allows the determination of the Prostate Imaging and could be used as a potential imaging biomarker. Reporting & Data System (PI-RADS) score [7], differ - Therefore, the present study aimed to explore the value entiation of malignant lesions from benign lesions, pre- of the APT-weighted MRI combined with serum PSA operative evaluation of malignant lesions, and evaluation levels to differentiate malignant prostate lesions from of treatment efficacy [8, 9]. Nevertheless, the sensitivity benign prostate lesions, and to provide more imaging and specificity of MRI for prostate cancer can still be information for early preoperative diagnosis of prostate improved [10, 11]. A recent study explored the diagnostic cancer. value of combination of amide proton transfer (APT) and mpMRI in transition zone (TZ) prostate cancer. Differ - Methods ences in APT-weighted and apparent diffusion coefficient Study design and patients (ADC) values between TZ prostate cancer and benign Data of patients who underwent prostate examinations prostatic hyperplasia (BPH), and differences in T2* values at Chongqing University Cancer Hospital (Chongqing, between stromal BPH and glandular BPH were found. It China) between July 2019 and March 2022 were ret- was found that APT-weighted and ADC were independ- rospectively analyzed. This study was approved by the ent predictors of TZ prostate cancer. Moreover, a combi- Ethics Committee of Chongqing University Cancer Hos- nation of APT and ADC values improved the diagnostic pital. The requirement for patients’ informed consent was sensitivity of TZ prostate cancer and achieved the pur- waived by the Ethics Committee. pose of improving the diagnostic efficiency [12]. All patients underwent T2-weighted imaging Novel MRI techniques are limited to conventional (T2WI), APT, diffusion-weighted imaging (DWI), and medical imaging, present the details of morphological dynamic contrast-enhanced (DCE) MRI. The inclu- features of lesions, and innovatively provide functional sion criteria were as follows: 1) untreated patients with Y ang et al. Cancer Imaging (2023) 23:3 Page 3 of 8 prostate diseases, and 2) patients who received MRI independent-samples t-test or the Wilcoxon rank-sum examination. The exclusion criteria were as follows: test. Two-sided P < 0.05 was considered statistically sig- 1) history of prostate surgery or endocrine therapy, nificant. Receiver operating characteristic (ROC) curves or 2) patients who underwent prostate biopsy within were plotted for the analysis. For parameters with statisti- 4–6 weeks before MRI. cally significant differences, the logistic regression analy - sis was used to estimate the probability of the combined MRI examination parameters. The parameters with statistically significant An INGENIA 3.0  T MRI scanner (Philips Healthcare differences were modeled using the binary logistic regres - Co., Ltd., Best, The Netherlands) was used for the MRI sion analysis, and the probability of the combined param- of prostate of all patients from July 2019 to March eters was calculated. ROC curves were plotted to compare 2022, including APT, DWI, T2WI, and DCE, using a the combinations of parameters in differentiating malig - 32-channel phased-array body coil (Table 1). The scan- nant prostate lesions from benign prostate lesions. ning parameters included b = 0 and 1400  s/mm for The index with the highest diagnostic accuracy was DWI and SPIR for fat suppression in APT. selected from the separate assessment of APT/ADC and PSA indexes to establish the combined model. The Image analysis SPSS software was employed to select the binary logis- The images were processed by the ISP post-processing tic regression model, and two or more parameters were workstation (Philips Healthcare Co., Ltd.). Two radi- obtained through modeling, in order to jointly predicting ologists reviewed the images independently and blindly. the probability of malignant tumors. The maximum lesion level was selected using the T2-weighted images as the standard. Then the region of Results interest (ROI) was selected at the same level on the APT General characteristics of patients and DWI images to measure the APT and ADC values A total of 123 patients were recruited, of whom 44 (Fig.  1). The averages of values measured by the two patients were excluded according to the exclusion cri- radiologists were calculated for the statistical analysis. teria. Finally, 79 patients were included in this study, of Patients’ medical records were reviewed to obtain the which 52 patients were in the malignant group (patholog- total PSA (tPSA) and free PSA (FPSA). ically proven with prostate cancer [20]) and 27 patients were in the benign group (pathologically proven with Statistical analysis non-cancerous lesions, including prostatic hyperplasia SPSS 23.0 software (IBM, Armonk, NY, US) was used and prostatitis). The mean age was 70.54 ± 8.66 years old. to perform the statistical analysis. Intraclass correlation There were no significant differences in age and volume (ICC) analysis was performed for the measurement data of prostate between malignant group and benign group obtained by the two radiologists (ICC > 0.75 indicated (P < 0.05) (Table 2). a high consistency, ICC equal to 0.4–0.75 indicated a moderate consistency, and ICC < 0.4 indicated a low con- Consistency between radiologists sistency). The Kolmogorov–Smirnov test was used to ICC consistency was performed for all the measurement assess the normal distribution, and the Levene’s test was data by the two radiologists. As shown in Table  3, APT employed for the homogeneity assessment of the con- and ADC parameters showed a moderate consistency tinuous data. Continuous data were analyzed using the between the two radiologists. Table 1 Parameters of the scanning sequences Sequences TR (ms) TE (ms) TA Layers Layer Interlayer FOV (mm) Matrix NSA (s) thickness spacing (mm) (mm) T2WI 3000 110 24 3 0.3 200 × 200 308 × 255 1 TSE factor 14 1′56’’ DWI 4828 90 3′8’’ 24 3 0 200 × 200 80 × 81 1 EPI factor 55 APT 5842 7.9 4′54’’ 18 5 0 140 × 140 80 × 78 1 TSE factor 174 T1WI 4.0 2.0 6′24’’ 24 3.5 0 250 × 250 180 × 140 1 dynamic phase 80 TR repetition time, TE echo time, TA acquisition time, FOV field of view, NSA number of signals per acquisition, T2WI T2-weighted imaging, DWI diffusion-weighted imaging, APT amide proton transfer, T1WI T1-weighted imaging TSE turbo spin echo, EPI echo-planar imaging Yang et al. Cancer Imaging (2023) 23:3 Page 4 of 8 Fig. 1 Case examples. 1 A 75-year-old man was hospitalized, and the physical examination revealed that PSA level increased. (1A) T2WI. (1B) APT. (1C) DWI. (1D) ADC. 2 A 68-year-old man was hospitalized for treatment after finding the possibility of prostate malignancy in the external hospital. (2A) T2WI. (2B) APT. (2C) DWI. (2D) ADC Ability of the quantitative parameters in differentiating malignant prostate lesions from benign prostate lesions Table 3 ICC consistency test of the results measured by the two radiologists As presented in Table  4, the quantitative parameters in differentiating malignant prostate lesions from benign Radiologist 1 Radiologist 2 ICC 95% CI prostate lesions were significantly different (P < 0.05), APTmean (%) 2.44 ± 1.21 2.42 ± 1.38 0.732 0.649–0.789 except for APTmin (P > 0.05). Quantitative APT param- APTmax (%) 5.56 ± 2.12 6.06 ± 2.26 0.701 0.605–0.768 eters, including APTmax and APTmean, could well dis- APTmin (%) -1.61 ± 2.33 -1.55 ± 2.62 0.652 0.538–0.734 tinguish malignant prostate lesions from benign prostate –3 ADCmean (10 0.91 ± 0.32 0.88 ± 0.37 0.664 0.554–0.743 lesions. Comparatively, APTmax had a higher sensitivity mm /s) (92.310), and APTmean had a higher specificity (77.780), –3 ADCmax (10 1.55 ± 0.45 1.47 ± 0.51 0.629 0.507–0.718 while they both had high area under the curve (AUC) val- 2 mm /s) ues (Fig.  1A and Table  5).The other parameters had rela - –3 ADCmin (10 0.53 ± 0.34 0.61 ± 0.53 0.566 0.395–0.699 tively high values in differentiating malignant prostate mm /s) lesions from benign prostate lesions. ROC curves were ICC intraclass correlation, CI confidential interval, APT amide proton transfer, ADC apparent diffusion coefficient Table 2 Characteristics of the patients Characteristics (n = 79) Malignant (n = 52) Benign (n = 27) P value Age (years) 69.29 ± 8.20 72.19 ± 9.43 0.42 Volume of prostate (cm ) 9.34 ± 12.46 10.83 ± 14.66 0.22 PSAD (ng/mL/cm ) 69.73 ± 137.59 3.76 ± 7.83 0.00 Gleason score, n (%) Low-risk subgroup (≤ 3 + 3 points) 15 (28.85%) High-risk subgroup (> 3 + 3 points) 37 (71.15%) PSAD prostate-specific antigen density Y ang et al. Cancer Imaging (2023) 23:3 Page 5 of 8 Table 4 Ability of the quantitative parameters in differentiating malignant from benign prostate diseases Malignant (n = 52) Benign (n = 27) Z/t P value APTmean (%) 2.740 ± 1.106 1.859 ± 1.200 -3.263 0.002 APTmax (%) 6.698 ± 1.837 4.530 ± 1.894 -4.924 < 0.001 APTmin (%) -1.788 ± 2.557 -1.259 ± 1.815 0.965 0.342 –3 2 ADCmean (10 mm /s) 0.845 (0.300–2.030) 1.190 (0.620–1.680) -4.259 < 0.001 –3 2 ADCmax (10 mm /s) 1.450 ± 0.445 1.738 ± 0.391 2.835 0.006 –3 2 ADCmin (10 mm /s) 0.464 ± 0.356 0.646 ± 0.271 2.330 0.022 tPSA (ng/mL) 33.604 (0.600–8169.000) 6.053 (0.840–131.760) -3.321 0.001 FPSA (ng/mL) 3.959 (0.020–483.000) 1.230 (0.210–7.240) -2.506 0.012 FPSA/tPSA 16.159 ± 15.300 23.336 ± 12.954 2.074 0.043 PSAD (ng/mL/cm ) 10.585 (0.010–688.540) 1.980 (0.030–38.650) -2.884 0.004 APT amide proton transfer, ADC apparent diffusion coefficient, tPSA total prostate-specific antigen, FPSA free prostate-specific antigen, PSAD prostate-specific antigen density Table 5 Analysis of the ROC curves of the different quantitative parameters AUC 95%CI Sensitivity Specificity Youden index J Best cut-off value P value APTmax 0.780 0.673–0.865 92.310 59.260 0.516 4.300 < 0.001 APTmean 0.710 0.598–0.807 63.460 77.780 0.412 2.300 0.001 ADCmean 0.793 0.688–0.876 75.000 81.480 0.565 1.060 < 0.001 ADCmax 0.689 0.575–0.788 46.150 92.580 0.388 1.360 0.002 ADCmin 0.689 0.575–0.789 73.080 70.370 0.435 0.620 0.003 tPSA 0.734 0.628–0.836 57.690 91.300 0.490 28.140 < 0.001 FPSA 0.684 0.565–0.788 42.000 100.000 0.420 7.240 0.003 FPSA/tPSA 0.684 0.565–0.788 68.000 69.570 0.376 17.050 0.007 PSAD 0.700 0.594–0.809 53.850 95.650 0.495 8.210 < 0.001 ROC receiver operating characteristic, AUC area under the curve, CI confidence interval, APT amide proton transfer, ADC apparent diffusion coefficient, tPSA total prostate-specific antigen, FPSA free prostate-specific antigen, PSAD prostate-specific antigen density DeLong et al., [21] plotted and showed that ADCmean had the highest AUC the combined parameters were significantly different value (0.793, 95% confidence interval (CI): 0.688–0.876), from the corresponding single parameter (P < 0.05). APTmax had the highest sensitivity (92.310), and FPSA had the highest specificity (100.000) (Fig.  2A, B, C and Discussion Table 5). Early diagnosis of prostate cancer improves its prognosis [1, 2], while screening tools should be further upgraded Combinations of parameters in differentiating malignant [10, 11, 14, 15]. Therefore, the present study aimed to prostate lesions from benign prostate lesions explore the value of the novel functional MRI technique, As reported previously, APTmax, APTmean, and APT-weighted MRI, combined with serum PSA levels for ADCmean [22], the most representative parameters of differentiating malignant prostate lesions from benign APT and ADC, were used in combination with PSAD, prostate lesions. The results indicated that APTmax, which is the most representative parameter of PSA APTmean, ADCmean, ADCmax, ADCmin, tPAD, FPSA, [23]. The combinations of these quantitative param- FPSA/tPSA, and PSAD had a high clinical value in differ - eters in differentiating malignant prostate lesions from entiating malignant prostate lesions from benign prostate benign prostate lesions were compared, and the ROC lesions. The combination of APTmax, APTmean, ADC - curves were plotted. As illustrated in Fig.  2D, E, F and mean, and PSAD showed the highest diagnostic value. Table 6, the AUC value was the highest when APTmax, Early diagnosis has important clinical significance for APTmean, and ADCmean were combined with PSAD the treatment and prognosis of patients with prostate (AUC: 0.880, 95% CI: 0.784–0.943, sensitivity: 86.540, cancer. Nevertheless, the early differentiation of malig - specificity: 78.260). According to the DeLong test [21], nant prostate lesions from benign prostate lesions is Yang et al. Cancer Imaging (2023) 23:3 Page 6 of 8 Fig. 2 Receiver operating characteristic (ROC) curves. A ROC curves of APT parameters. B ROC curves of the DWI parameters. C ROC curves of the PSA parameters. D ROC curves of the combinations of APTmax, ADCmean, and PSAD. D ROC curves of the combinations of APTmean, ADCmean, and PSAD. E ROC curves of the combinations of APTmax, APTmean, ADCmean, and PSAD. APT: amide proton transfer; ADC: apparent diffusion coefficient; tPSA: total prostate-specific antigen; FPSA: free prostate-specific antigen; ratio: FPSA/tPSA. PSAD: prostate-specific antigen density Table 6 Analysis of the ROC curves of the combinations of different quantitative parameters AUC 95%CI Sensitivity Specificity Youden index J P value APTmax + ADCmean 0.841 0.763–0.928 92.310 70.370 0.627 < 0.001 APTmax + PSAD 0.826 0.721–0.904 80.770 69.570 0.503 < 0.001 ADCmean + PSAD 0.854 0.753–0.925 75.000 91.300 0.663 < 0.001 APTmean + ADCmean 0.837 0.748–0.918 78.850 85.190 0.640 < 0.001 APTmean + PSAD 0.826 0.721–0.904 75.000 82.960 0.576 < 0.001 APTmean + ADCmean + PSAD 0.878 0.782–0.942 80.770 91.300 0.721 < 0.001 APTmax + APTmean + ADCmean + PSAD 0.880 0.784–0.943 86.540 78.260 0.648 < 0.001 ROC receiver operating characteristic, AUC area under the curve, CI confidence interval, APT amide proton transfer, ADC apparent diffusion coefficient, tPSA total prostate-specific antigen, FPSA free prostate-specific antigen, PSAD prostate-specific antigen density DeLong et al., [21] still difficult in clinical practice based only on the cur - signal values, depending on the exchange ratio rent imaging methods, such as ultrasound and MRI. between the amide protons and free-water protons. The APT imaging technique is based on transferring The exchange ratio depends on pH values and protein the amide protons and water, and it reflects the changes concentrations in the body. The APT technique was of proteins and pH values by variations of water sig- initially used for the nervous system [24–26]. In recent nals. The internal contrast is acquired by measuring the years, a great number of researchers have applied the water signals to indirectly acquire the APT-weighted APT technique to diagnose prostate diseases [27, 28]. Y ang et al. Cancer Imaging (2023) 23:3 Page 7 of 8 The findings of the present study showed that APTmax respectively) [29]. Guo et al. [12] found that APTmean and had a high diagnostic value for differentiating malignant ADC were independent predictors of TZ prostate cancer. prostate lesions from benign prostate lesions. The sen - Moreover, combination of APTmean and ADC values sitivity was the highest, indicating that the maximum improved the sensitivity of the diagnosis of TZ prostate transfer of amide protons and exchange ratio of water cancer and achieved the purpose of improving the diag- protons in a lesion could sensitively reflect the occur - nostic efficiency, which are similar to the results of the pre - rence of malignant lesions. Jia et al. [27], for the first time, sent study. attempted to apply the APT imaging in prostate diseases, The advantage of APT-weighted MRI is that it is a 3D and reported the value of this technique in differentiating imaging technique. Compared with the conventional malignant prostate lesions from benign prostate lesions. two-dimensional (2D) APT technique [30], this tech- Takayama et  al. [28] found that the APT-weighted val- nique could scan multiple layers in a short time, acquire ues in prostate cancer patients with a Gleason score of the APT image of the whole prostate region, and provide 7 points were significantly higher than those of patients more comprehensive functional information. with other scores. These findings were generally in agree - There were several limitations in the present study. ment with our results. The metabolism in malignant First, the sample size was small and imbalance, and addi- prostate lesions is more active than in benign lesions. The tional studies are required to determine the exact diag- exchange ratio of protons is higher, which is consistent nostic value of APT for prostate cancer. Second, this was with the pathological features of malignant lesions. The a single-center study, and local practice biases could influ - differences of APTmax and APTmean, two parameters ence the results. Multicenter studies can not only increase acquired by APT imaging, were statistically significant. the sample size, but also mitigate the risk of bias. Last but In contrast, the difference of APTmin was not statistically not least, because of the small sample size, no direct com- significant, which could be associated with the fact that parison was performed among imaging techniques. APTmin expresses the lowest value of protein content in the region of interest, and the difference is not enough Conclusions to distinguish between benign and malignant lesions. APT has a diagnostic value for prostate cancer. APTmax, Importantly, the minimum exchange ratio of protons APTmean, ADCmean, ADCmin, tPAD, FPSA, and PSAD could not reflect the degree of metabolic activity. showed to have a high diagnostic value in differentiating The results of the present study revealed that the differ - malignant prostate lesions from benign prostate lesions. ences of APTmax, APTmean, ADCmean, ADCmax, ADC- The combination of APTmax, APTmean, ADCmean, and min, tPAD, FPSA, FPSA/tPSA, and PSAD were statistically PSAD had the highest diagnostic value. significant (P < 0.05), suggesting that these parameters had high diagnostic values in differentiating malignant prostate Abbreviations lesions from benign prostate lesions. Among these param- ADC Apparent diffusion coefficient eters, ADCmean had the highest AUC, APTmax had the APT Amide proton transfer AUC Ar ea under the curve highest sensitivity, and FPSA had the highest specificity CEST Chemical exchange saturation transfer and the lowest sensitivity, reflecting that FPSA had the DWI Diffusion-weighted imaging highest diagnostic accuracy and a relatively high false-neg- FPSA Free prostate-specific antigen ICC Intraclass correlation ative rate. This indicated that functional MRI sequences, MRI Magnetic resonance imaging such as DWI and PSA, can be used as independent pre- PI-RADS P rostate Imaging Reporting & Data System dictive biomarkers to discriminate benign prostate lesions PSA Prostate-specific antigen ROC Receiver operating characteristic and malignant prostate lesions. In contrast, APTmax had ROI Region of interest the highest positive rate. After combining these param- DCE T1-weighted dynamic contrast-enhanced (DCE) eters, the results showed that the combination of APTmax, T2WI T2-weighted imaging tPSA Total prostate-specific antigen APTmean, ADCmean, and PSAD had the highest AUC and sensitivity, suggesting that this combination had the Acknowledgements highest diagnostic value for prostate cancer. Compared The authors would like to thank all participants who were enrolled in this study. We highly appreciate the efforts dedicated by Dr. Xiaoyong Zhang from with using PSA and DWI alone, the combination of APT, Philips Healthcare (Clinical Science) for his assistance in APT-MRI sequence DWI, and PSA techniques had a relatively high diagnostic development and valuable discussions. value (AUC: 0.880) and a high sensitivity (86.540) for pros- Authors’ contributions tate cancer. A previous study demonstrated that the AUC Conceptualization: Jiuquan Zhang; Data curation: Yuchuan Tan; Formal values for the biparametric MRI (bpMRI) and multipara- analysis: Hanli Dan; Funding acquisition: Lu Yang; Investigation: Yipeng Zhang; metric MRI (mpMRI) protocols for prostate cancer were Methodology: Peng Xian; Project administration: Lu Yang; Resources: Meng Lin; Software: Yong Tan; Supervision: Jiuquan Zhang; Validation: Lu Yang; Roles/ comparable (0.790 [0.732–0.840] and 0.791 [0.733–0.841], Yang et al. Cancer Imaging (2023) 23:3 Page 8 of 8 Writing—original draft: Lei Wang; Writing—review and editing: Lu Yang. The 10. Zhen L, Liu X, Yegang C, Yongjiao Y, Yawei X, Jiaqi K, et al. Accuracy of mul- author(s) read and approved the final manuscript. tiparametric magnetic resonance imaging for diagnosing prostate Can- cer: a systematic review and meta-analysis. BMC Cancer. 2019;19(1):1244. Funding 11. Becker AS, Kirchner J, Sartoretti T, Ghafoor S, Woo S, Suh CH, et al. Interac- This study was supported by the National Natural Science Foundation of tive, up-to-date meta-analysis of MRI in the management of men with China (Grant No. 82071883), the General Project of Chongqing Natural Science suspected prostate cancer. J Digit Imaging. 2020;33(3):586–94. Foundation (Grant No. cstc2021jcyj-msxmX0447), the China National Cancer 12. Guo Z, Qin X, Mu R, Lv J, Meng Z, Zheng W, et al. Amide proton transfer Center Climbing Fund (Grant No. NCC201822872), and the Open Research could provide more accurate lesion characterization in the transition Fund Project of Chongqing University Cancer Hospital (2020). The funder had zone of the prostate. J Magn Reson Imaging. 2022;56(5):1311–9. the following roles in the study: study design, collection, analysis, and interpre- 13. Pezaro C, Woo HH, Davis ID. Prostate cancer: measuring PSA. Intern Med J. tation of data, writing the paper and decision to submit it for publication. 2014;44(5):433–40. 14. Ilic D, Djulbegovic M, Jung JH, Hwang EC, Zhou Q, Cleves A, et al. Prostate Availability of data and materials cancer screening with prostate-specific antigen (PSA) test: a systematic All data generated or analyzed during this study are included in this published review and meta-analysis. BMJ. 2018;362:k3519. article. 15. Lee SK, Kim J-Y, Jeong HS. Benign peripheral nerve sheath tumor of digit versus major-nerve: comparison of MRI findings. PLoS ONE. 2020;15(3):e0230816. Declarations 16. Meng N, Wang X, Sun J, Han D, Ma X, Wang K, et al. Application of the amide proton transfer-weighted imaging and diffusion kurtosis imaging Ethics approval and consent to participate in the study of cervical cancer. Eur Radiol. 2020;30(10):5758–67. This study was approved by the Ethics Committee of Chongqing Univer- 17. Kamitani T, Sagiyama K, Togao O, Yamasaki Y, Hida T, Matsuura Y, et al. sity Cancer Hospital, and it was performed in accordance with the ethical Amide proton transfer (APT ) imaging of parotid tumors: Differentiation of standards laid down in an appropriate version of the Declaration of Helsinki malignant and benign tumors. Eur J Radiol. 2020;129:109047. (as revised in Brazil 2013). The requirement for patients’ informed consent was 18. Ma X, Bai Y, Lin Y, Hong X, Liu T, Ma L, et al. Amide proton transfer mag- waived by the Ethics Committee. netic resonance imaging in detecting intracranial hemorrhage at differ - ent stages: a comparative study with susceptibility weighted imaging. Sci Consent for publication Rep. 2017;7:45696. Not applicable. 19. Sartoretti T, Sartoretti E, Wyss M, Schwenk A, Najafi A, Binkert C, et al. Amide proton transfer contrast distribution in different brain regions in Competing interests young healthy subjects. Front Neurosci. 2019;13:520. The authors declare that they have no competing interests. 20. Humphrey PA, Moch H, Cubilla AL, Ulbright TM, Reuter VE. The 21. DeLong, E, DeLong, D, Clarke-Pearson, D. Comparing the Areas under Author details Two or More Correlated Receiver Operating Characteristic Curves: A Department of Radiology, Chongqing University Cancer Hospital & Nonparametric Approach BIOMETRICS. 1988;44(3):837. Chongqing Cancer Institute & Chongqing Cancer Hospital, No.181 Hanyu 22. Saito S, Koyama Y, Ueda J, Hashido T. Relationship between apparent Road, Shapingba District, Chongqing 400030, China. Department of Urol- diffusion coefficient distribution and cancer grade in prostate cancer and ogy, Chongqing University Cancer Hospital & Chongqing Cancer Institute & benign prostatic hyperplasia. Diagnostics (Basel). 2022;12(2):525. Chongqing Cancer Hospital, Chongqing 400030, China. 23. Milkovic B, Dzamic Z, Pejcic T, Kajmakovic B, Nikolic D, Cirovic D, et al. Evaluation of free-to-total prostate specific antigen (F/T PSA), prostate Received: 21 March 2022 Accepted: 19 December 2022 specific antigen density (PSAD) and (F/T )/PSAD sensitivity on reduction of unnecessary prostate biopsies for patients with PSA in gray zone. Ann Ital Chir. 2014;85(5):448–53. 24. Kang XW, Xi YB, Liu TT, Wang N, Zhu YQ, Wang XR, et al. Grading of Glioma: combined diagnostic value of amide proton transfer weighted, References arterial spin labeling and diffusion weighted magnetic resonance imag- 1. Mottet N, Corford P, van den Bergh RCN, Briers E, De Santis M, Fanti S, ing. BMC Med Imaging. 2020;20(1):50. et al. European Association of Urology (EAU). Guidelines on prostate 25. Zhang Z, Zhang C, Yao J, Chen X, Gao F, Jiang S, et al. Protein-based cancer. Arnhem: EAU Guidelines Office; 2020. amide proton transfer-weighted MR imaging of amnestic mild cognitive 2. NCCN Clinical Practice Guidelines in Oncology (NCCN Guidelines). Pros- impairment. Neuroimage Clin. 2020;25:102153. tate Cancer. Version 3.2020. Fort Washington: National Comprehensive 26. Zhou J, Heo HY, Knutsson L, van Zijl PCM, Jiang S. APT-weighted MRI: Cancer Network; 2020. Techniques, current neuro applications, and challenging issues. J Magn 3. Bray F, Ferlay J, Soerjomataram I, Siegel RL, Torre LA, Jemal A. Global Reson Imaging. 2019;50(2):347–64. cancer statistics 2018: GLOBOCAN estimates of incidence and mor- 27. Jia G, Abaza R, Williams JD, Zynger DL, Zhou J, Shah ZK, et al. Amide pro- tality worldwide for 36 cancers in 185 countries. CA Cancer J Clin. ton transfer MR imaging of prostate cancer: a preliminary study. J Magn 2018;68(6):394–424. Reson Imaging. 2011;33(3):647–54. 4. Jemal A, Bray F, Center MM, Ferlay J, Ward E, Forman D. Global cancer 28. Takayama Y, Nishie A, Sugimoto M, Togao O, Asayama Y, Ishigami K, et al. statistics. CA Cancer J Clin. 2011;61(2):69–90. Amide proton transfer (APT ) magnetic resonance imaging of prostate 5. Siegel RL, Miller KD, Jemal A. Cancer statistics, 2020. CA Cancer J Clin. cancer: comparison with Gleason scores. MAGMA. 2016;29(4):671–9. 2020;70(1):7–30. 29. Xu L, Zhang G, Shi B, Liu Y, Zou T, Yan W, et al. Comparison of biparametric 6. Nyame YA, Gore JL. What goes up must come down: identifying truth and multiparametric MRI in the diagnosis of prostate cancer. Cancer from global prostate cancer epidemiology. Eur Urol. 2020;77(1):53–4. Imaging. 2019;19(1):90. 7. PI-RADS. Prostate Imaging - Reporting and Data System Version 21. 30. Nishie A, Takayama Y, Asayama Y, Ishigami K, Ushijima Y, Okamoto D, et al. Philadelphia: American College of Radiology; 2019. Amide proton transfer imaging can predict tumor grade in rectal cancer. 8. Winkel DJ, Wetterauer C, Matthias MO, Lou B, Shi B, Kamen A, et al. Magn Reson Imaging. 2018;51:96–103. Autonomous detection and classification of PI-RADS lesions in an MRI screening population incorporating multicenter-labeled deep learn- Publisher’s Note ing and biparametric imaging: proof of concept. Diagnostics (Basel). Springer Nature remains neutral with regard to jurisdictional claims in pub- 2020;10(11):951. lished maps and institutional affiliations. 9. Qi Z, Li W, Tan J, Wang C, Lin H, Zhou B, et al. Eec ff t of ginsenoside Rh(2) on renal apoptosis in cisplatin-induced nephrotoxicity in vivo. Phy- tomedicine. 2019;61:152862.

Journal

Cancer Imaging – Springer Journals

Published: Jan 7, 2023

Keywords: Amide proton transfer (APT)-weighted MRI; Prostate-specific antigen; Prostate cancer; Malignant lesions; Benign lesions

There are no references for this article.