Get 20M+ Full-Text Papers For Less Than $1.50/day. Subscribe now for You or Your Team.

Learn More →

Esophageal stenosis and the Glasgow Prognostic Score as independent factors of poor prognosis for patients with locally advanced unresectable esophageal cancer treated with chemoradiotherapy (exploratory analysis of JCOG0303)

Esophageal stenosis and the Glasgow Prognostic Score as independent factors of poor prognosis for... Int J Clin Oncol (2017) 22:1042–1049 DOI 10.1007/s10147-017-1154-6 ORIGINAL ARTICLE Esophageal stenosis and the Glasgow Prognostic Score as independent factors of poor prognosis for patients with locally advanced unresectable esophageal cancer treated with chemoradiotherapy (exploratory analysis of JCOG0303) 1 2 3 4 2 Tatsuya Okuno  · Masashi Wakabayashi  · Ken Kato  · Masayuki Shinoda  · Hiroshi Katayama  · 5 6 7 8 9 10 Hiroyasu Igaki  · Yasuhiro Tsubosa  · Takashi Kojima  · Hiroshi Okabe  · Yusuke Kimura  · Tatsuyuki Kawano  · 11 12 5 2 2 13 Shinichi Kosugi  · Yasushi Toh  · Hoichi Kato  · Kenichi Nakamura  · Haruhiko Fukuda  · Satoshi Ishikura  · 14 15 Nobutoshi Ando  · Yuko Kitagawa  · Japan Esophageal Oncology Group/Japan Clinical Oncology Group Received: 28 December 2016 / Accepted: 14 June 2017 / Published online: 17 July 2017 © The Author(s) 2017. This article is an open access publication Abstract CRP ≤1.0  mg/dL and albumin ≥3.5  g/dL were classified Background The aim of this study was to investigate the as GPS0. If only CRP was increased or only albumin was possible prognostic factors and predictive accuracy of the decreased, the patients were classified as GPS1, and the Glasgow Prognostic Score (GPS) for patients with unre- patients with CRP >1.0  mg/dL and albumin <3.5  g/dL sectable locally advanced esophageal squamous cell carci- were classified as GPS2. noma (LAESCC) treated with chemoradiotherapy. Results The patients’ backgrounds were as follows: median Methods One hundred forty-two patients were enrolled age (range), 62 (37–75); male/female, 119/12; ECOG PS in JCOG0303 and assigned to the standard cisplatin and 0/1/2, 64/65/2; and clinical stage (UICC 5th) IIB/III/IVA/ 5-fluorouracil (PF)-radiotherapy (RT) group or the low- IVB, 3/75/22/31. Multivariable analyses indicated only dose PF-RT group. One hundred thirty-one patients with esophageal stenosis as a common factor for poor prognosis. sufficient data were included in this analysis. A Cox regres - In addition, overall survival tended to decrease according to sion model was used to analyze the prognostic factors of the GPS subgroups (median survival time (months): GPS0/ patients with unresectable LAESCC treated with PF-RT. GPS1/GPS2 16.1/14.9/8.7). The GPS was classified based on the baseline C-reactive Conclusions Esophageal stenosis was identified as a protein (CRP) and serum albumin levels. Patients with candidate stratification factor for randomized trials of * Tatsuya Okuno Department of Surgery, Kyoto University Graduate School t-okuno@pg7.so-net.ne.jp of Medicine, Kyoto, Japan Department of Surgery, Iwate Medical University School Division of Gastroenterology, Department of Internal of Medicine, Morioka, Japan Medicine, Graduate School of Medicine, Kobe University, 7-5-1 Kusunoki Chuo, Kobe 650-0017, Japan Tokyo Medical and Dental University, Tokyo, Japan Japan Clinical Oncology Group Data Center/Operations Department of Surgery, Niigata University Medical Office, National Cancer Center, Tokyo, Japan and Dental Hospital, Niigata, Japan 3 12 Gastrointestinal Medical Oncology Division, National Department of Gastroenterological Surgery, National Kyushu Cancer Center Hospital, Tokyo, Japan Cancer Center, Fukuoka, Japan 4 13 Department of Gastrointestinal Surgery, Aichi Cancer Center Department of Radiology, Koshigaya Municipal Hospital, Hospital, Nagoya, Japan Koshigaya, Saitama, Japan 5 14 Esophageal Surgery Division, National Cancer Center Department of Surgery, International Goodwill Hospital, Hospital, Tokyo, Japan Yokohama, Japan 6 15 Division of Esophageal Surgery, Shizuoka Cancer Center Department of Surgery, Keio University, School of Medicine, Hospital, Shizuoka, Japan Tokyo, Japan Gastrointestinal Oncology Division, National Cancer Center Hospital East, Kashiwa, Japan Vol:.(1234567890) 1 3 Int J Clin Oncol (2017) 22:1042–1049 1043 unresectable LAESCC patients. Furthermore, GPS repre- in therapeutic effectiveness and prognosis is important sents a prognostic factor for LAESCC patients treated with for tailoring an optimized treatment strategy and further chemoradiotherapy. improving treatment outcomes. Emerging evidence sug- Clinical Trial Information UMIN000000861. gests that TNM stage, a weight loss of more than 10% of one’s body mass, dysphagia, large tumors, older age, Keywords T4 esophageal cancer · Chemoradiotherapy · and lymphatic micrometastases (identified by immuno- Prognostic factor · Esophageal stenosis · Glasgow histochemical analysis) are independent predictors of a Prognostic Score poor prognosis in patients with advanced EC who have undergone surgery [1]. However, little is known about the prognostic factors in patients with advanced unresectable Introduction LAESCC who have been treated with CRT. Therefore, we investigated clinical pretreatment factors that might affect Despite substantial improvements in screening and diag- the survival of patients enrolled in the JCOG0303 trial. nosis, esophageal cancer (EC) is frequently diagnosed at Moreover, hypoalbuminemia and C-reactive protein a very advanced stage [1]. The thoracic esophagus lacks (CRP) are associated with a poor prognosis in cancer serosa and is closely surrounded by the trachea, bronchus, patients [10]. The Glasgow Prognostic Score (GPS) com- lung and aorta. Thus, esophageal cancer is more likely bines albumin and CRP into a risk stratification score to invade these vital organs and become unresectable. for the prognosis of clinical outcome in cancer patients. According to the Comprehensive Registry of Esophageal This scoring system has been validated in colorectal can- Cancer in Japan, the incidence of cT4 esophageal carci- cer (CRC) and other malignant tumors, including EC noma is approximately 15% among all EC patients [2]. [11–15]. Given that hypoalbuminemia and CRP are asso- Curative resection is not feasible in patients with locally ciated with poor prognosis in cancer patients [10], we advanced esophageal squamous cell carcinoma (LAESCC) also evaluated the potential prognostic role of the GPS on if the cancer has invaded other organs, and such cases long-term outcome in patients undergoing definitive CRT have an unfavorable prognosis [3–6]. Definitive chemora- for LAESCC. diotherapy (CRT) is the standard of care that is currently available for unresectable locally advanced EC [5–8]. In a former phase II trial, 18 of 54 (33%) patients with clini- cal T4 and/or M1 lymph nodes (M1Lym) and esophageal Methods squamous cell carcinoma (ESCC), who received cisplatin and 5-fluorouracil (PF) with concurrent 60-Gy irradia- Schema of the JCOG0303 Study tion, achieved a complete response (CR); in that study, the median overall survival (OS) and the 3-year survival rate Key eligibility criteria were as follows: age of 75  years were 9  months and 23%, respectively [7]. Additionally, a or younger, patients with histologically proven squamous new multicenter trial JCOG0303 was conducted to evalu- cell carcinoma, adenosquamous, or basaloid carcinoma ate the efficacy and toxicity, and particularly the long-term of the thoracic esophagus, and an Eastern Cooperative outcome, of CRT in patients with unresectable LAESCC. Oncology Group (ECOG) performance status (PS) of The aim of JCOG0303 was to evaluate whether RT plus 0–2; in addition, patients who had any of the following low-dose cisplatin and 5-fluorouracil (LDPF) chemother - conditions were included: definite clinical T4 cancer, at apy had an advantage in terms of survival and/or toxicity least 1 unresectable metastatic regional lymph node due over RT plus standard-dose cisplatin and 5-fluorouracil to invasion into an adjacent organ, or computed tomog- (SDPF) chemotherapy in a randomized phase II/III trial. raphy (CT)-based evidence of M1 Lym disease, such as In a primary analysis of the phase II portion, the study fixed supraclavicular or celiac lymph nodes. Regional was terminated due to futility. In an updated analysis, the lymph nodes were defined using criteria specified by the median OS and the 3-year OS rate were 13.1 months and 5th edition of the Union for International Cancer Control 25.9%, respectively, in the SDPF-RT arm, and 14.4 months (UICC) TNM staging system [16]. After confirmation and 25.7%, respectively, in the LDPF-RT arm. The pri- of eligibility, the patients were randomized at the JCOG mary endpoint of OS was nearly equivalent in both treat- Data Center. Written informed consent was obtained ment arms [9]. from all enrolled patients, and the institutional review The status of each patient with invasive thoracic esoph- boards of all participating institutions approved the study ageal cancer varies widely due to variability in the extent protocol. The JCOG0303 trial was registered with the of the tumors and the nutritional and general status of the UMIN Clinic Trials Registry (http://www.umin.ac.jp/ctr/) patients. Thus, establishing the factors that are involved with the identification number UMIN000000861. 1 3 1044 Int J Clin Oncol (2017) 22:1042–1049 Treatment patients whose CRP level was >1.0 mg/dL and whose albu- min level was <3.5 g/dL were classified as GPS2. The patients were randomized to receive either SDPF-RT (arm A) or LDPF-RT (arm B). The chemotherapy regimen in arm A consisted of 70  mg⁄m cisplatin given on days 1 Results and 29 combined with a continuous infusion of 700 mg⁄m 5-FU given on days 1–4 and 29–32. Patients in arm B Patient characteristics received a 1  h infusion of 4  mg⁄m cisplatin before RT, combined with a continuous infusion of 200  mg⁄m 5-FU In all, 142 patients from 24 participating institutions of on the first 5 days of each week. Treatment completion was the Japan Esophageal Oncology Group of the JCOG were defined as the termination of two courses of chemotherapy enrolled between May 2000 and May 2006. Of these, 71 and 60 Gy of radiotherapy within 63 days. patients were randomly assigned to SDPF-RT (arm A) and 71 patients were randomly assigned to LDPF-RT (arm B). Statistical analysis Of the patients randomized into the two groups, 7 in arm A and 4 in arm B were excluded from this ancillary study For the analysis of prognostic factors using pretreatment because there were no documented data for the aforemen- factors, the OS was calculated from the date of randomiza- tioned prognostic factors. Thus, 131 patients comprised tion until death from any cause, or the OS was censored at the population for the analysis based on prognostic fac- the time of the last follow-up. An initial univariable Cox tors (Fig.  1). The patient characteristics in each group that regression analysis was performed to evaluate factors that was subjected to the analysis using pretreatment factors are could potentially affect the survival outcomes. Further - shown in Table 1. more, a multivariable analysis that included all covariates The median overall survival times (MSTs) of all rand- and a multivariable analysis with the stepwise selection omized 142 patients and of the final 131 patients included method were performed. in this study were 14.4  months. Among all randomized The following pretreatment factors were tested in the patients, the 1- and 3-year survival rates were 57.0 and analyses: age (≥65 vs. <65  years), sex (male vs. female), 23.9%, respectively, while those rates of the patients ECOG PS (0 vs. 1 or 2), clinical T stage (T1, T2, T3 vs. included in this study were 58.8 and 22.9%, respectively T4), clinical N stage (N0 vs. N1), clinical M stage (M0 vs. (Fig. 2). M1a or M1b), histological subtype of SCC (well-differen- tiated or moderately differentiated vs. poorly differentiated vs. unknown), location of the primary tumor (upper vs. All registered patients middle or lower), clinically diagnosed esophageal steno- (n = 142) sis (absent vs. present), adjacent organ invasion mediated by LN metastasis (absent vs. present), intramural metasta- SDPF-RT: 71 sis (absent vs. present), and the laboratory data described LDPF-RT: 71 below, which were obtained upon enrollment in the trial. The white blood cell count, neutrophil count, hemoglobin level (Hb), CRP level, albumin level, GPT, BUN, cre- Exclusion of 11 patients atinine clearance (CrC), CEA level, and SCC level were dichotomized, with cutoff points at 10,000/mm , 8000/ with missing data forat mm , 13  g/dL, 1.0  mg/dL, 3.5  g/dL, 30  IU/L, 15  mg/dL, least 1 baseline factor 80 mL/min, 5 ng/mL, and 1.5 ng/mL, respectively. The cut- Phase II portion (n = 2) off points of each laboratory test were determined based on a clinical perspective. Survival curves were estimated using the Kaplan–Meier method. Two-sided p values <0.05 were All analyzed patients considered statistically significant. All statistical analyses (n = 131) were performed with SAS9.2 (SAS Institute, Cary, NC, USA). SDPF-RT: 64 The GPS was determined as previously described by LDPF-RT: 67 several groups [12–15]. Patients with a CRP level ≤1.0 mg/ dL and an albumin level ≥3.5 g/dL were classified as GPS0. If only the CRP level was increased or if only the albumin level was decreased, patients were classified as GPS1, and Fig. 1 Patient flowchart 1 3 Int J Clin Oncol (2017) 22:1042–1049 1045 Table 1 Baseline patient characteristics (n = 131) histological subtype of SCC (unknown) were significantly associated with poor survival. A multivariable analysis Characteristic SDPF-RT (n = 64) LDPF-RT (n = 67) with stepwise selection also confirmed esophageal stenosis Gender as an indicator of poor survival, although a serum Alb level  Male 61 58 <3.5 g/dL was not an independent prognostic factor in the  Female 3 9 multivariable analysis that included all covariates. The haz- Age (years) ard ratio for patients with esophageal stenosis (present) was  Median 63 62 1.60 (95% CI 1.10–2.32) compared with patients without  Range 37–75 43–74 esophageal stenosis (absent) (p = 0.014) in the multivaria- ECOG PS ble analysis with stepwise selection (Table 2). Patients with  0 30 34 esophageal stenosis had a median survival of 9.9  months,  1 33 32 while patients without esophageal stenosis had a median  2 1 1 survival of 16.0 months (p = 0.001) (Fig. 3). cTNM (UICC 5th edition)  T1 2 0 GPS as an independent prognostic factor for survival  T2 5 2  T3 9 12 The association between survival and the GPS in patients  T4 48 53 with esophageal carcinoma is shown in Fig.  4. The HR  N0 9 9 for overall survival in GPS1 patients to GPS0 patients  N1 55 58 was 1.22 (95% CI 0.80–1.86; p  =  0.35) and was 1.95  M0 38 40 (95% CI 1.19–3.18; p  =  0.008) in GPS2 patients to GPS0  M1a 8 14 patients. Patients with a GPS of 0 had a median survival  M1b 18 13 of 16.1 months, while patients with a GPS of 1 and 2 had Clinical stage (UICC) a median survival of 14.9 and 8.7  months, respectively  IIb 2 1 (p = 0.001) (Fig. 4). These findings suggest that an elevated  III 36 39 GPS is associated with poor prognosis. Moreover, we con-  IVa 8 14 firmed that the presence or absence of stenosis and the GPS  IVb 18 13 score did not have remarkable relationship. ECOG PS Eastern Cooperative Oncology Group performance status Discussion 1.0 Here, we retrospectively investigated the therapeutic out- 0.9 comes of patients who were prospectively registered in the 0.8 JCOG0303 trial and evaluated pretreatment clinical param- 0.7 All randomized 0.6 Pts included eters that might consistently affect survival. In the present 0.5 HR = 1.08 (95%CI 0.55 –2.13) study, two multivariable analyses revealed esophageal 0.4 0.3 stenosis as the only single factor that was associated with 0.2 poor prognosis in patients with unresectable LAESCC. To 0.1 our knowledge, this is the first study to report an associa- 0.0 01234567 8 tion between treatment outcome and clinically diagnosed Years after randomization No. at Risk All randomized 142 81 47 34 23 18 11 5 0 esophageal stenosis. Pts included 131 77 43 30 20 16 9 4 0 Currently, the established factors of a poor prognosis are PS, tumor size and TNM stage, but in our analysis, no correlations were found between these factors and prog- Fig. 2 Kaplan–Meier estimates of the overall survival of all rand- nosis. This might reflect the fact that our patients in the omized (n = 142) patients, including the study (n = 131) patients present study were limited to those with T4 and/or unre- sectable supraclavicular lymph node metastasis (i.e., those Multivariable analysis at a relatively advanced stage with a poor prognosis). Zenda et  al. described that low pretreatment levels of Hb A multivariable analysis that included all covariates dem- were significantly associated with the efficacy of CRT in onstrated that esophageal stenosis, lower Hb, clinical M1 patients with T4/M1 EC [17]. Similarly, low levels of Hb stage, location of the primary tumor (middle or lower) and have been reported to be associated with sensitivity to 1 3 OverallSurvival 1046 Int J Clin Oncol (2017) 22:1042–1049 Table 2 Multivariable Cox regression analysis of overall survival according to pretreatment factors Factors Levels N Univariable analysis Multivariable analysis Including all covariates Stepwise selection HR (95% CI) p value HR (95% CI) p value HR (95% CI) p value Age 65> 85 1 0.320 1 0.231 65≤ 46 1.219 (0.825–1.802) 1.335 (0.832–2.140) Sex Female 12 1 0.464 1 0.744 Male 119 0.784 (0.409–1.504) 0.881 (0.414–1.878) PS 0 64 1 0.446 1 0.360 1 or 2 67 1.156 (0.796–1.678) 1.228 (0.791–1.908) T-stage T4 101 1 0.558 1 0.177 T1, T2, T3 30 0.873 (0.555–1.373) 0.646 (0.343–1.217) N-stage N0 18 1 0.348 1 0.976 N1 113 1.308 (0.747–2.292) 1.011 (0.512–1.996) M-stage M0 78 1 0.070 1 0.002 M1a or M1b 53 1.426 (0.972–2.092) 2.271 (1.368–3.772) Location of the pri- Ut 37 1 0.396 1 0.009 mary tumor Mt or Lt 94 1.199 (0.788–1.825) 2.068 (1.198–3.569) Histological subtype Well or Moderately 97 1 All (0.372) 1 All (0.024) of SCC differentiated Poorly differentiated 19 1.138 (0.680–1.903) 0.623 1.311 (0.741–2.320) 0.352 Unknown 15 1.494 (0.844–2.644) 0.168 2.844 (1.332–6.070) 0.007 Stenosis Absent 70 1 0.010 1 0.002 1 0.014 Present 61 1.629 (1.123–2.361) 2.014 (1.293–3.137) 1.598 (1.101–2.319) Adjacent organ Inva- Absent 105 1 0.822 1 0.545 sion mediated by LN Present 26 1.056 (0.657–1.698) 1.193 (0.675–2.108) metastasis Intramural metastasis Absent 107 1 All (0.081) 1 All (0.216) Present 13 1.641 (0.913–2.948) 0.098 1.883 (0.927–3.824) 0.080 Unknown 11 1.735 (0.922–3.266) 0.088 1.057 (0.456–2.451) 0.897 WBC 10,000≥ 113 1 0.611 1 0.095 10000< 18 0.868 (0.503–1.498) 0.449 (0.176–1.148) ANC 8000≥ 121 1 0.920 1 0.533 8000< 10 1.036 (0.520–2.065) 1.477 (0.434–5.027) Hb 13≥ 85 1 0.042 1 0.008 13< 46 0.664 (0.447–0.986) 0.509 (0.310–0.836) CRP 1.0 (mg/dL)≥ 69 1 0.117 1 0.367 1.0 (mg/dL)< 62 1.344 (0.929–1.945) 1.266 (0.758–2.114) Alb 3.5 (g/dL)> 40 1 0.018 1 0.988 1 0.025 3.5 (g/dL)≤ 91 0.621 (0.420–0.921) 0.996 (0.559–1.773) 0.636(0.429–0.944) GPT 30> 105 1 0.032 1 0.134 30≤ 26 1.639 (1.044–2.573) 1.531 (0.878–2.670) BUN 15.0> 98 1 0.863 1 0.336 15.0≤ 33 0.963 (0.628–1.478) 1.269 (0.781–2.063) CrC 80.0> 65 1 0.405 1 0.077 80.0≤ 66 1.171 (0.807–1.698) 1.500 (0.957–2.352) CEA 5> 110 1 0.849 1 0.733 5≤ 21 0.954 (0.587–1.550) 0.902 (0.501–1.626) SCC 1.5≥ 67 1 0.265 1 0.622 1.5< 64 1.236 (0.852–1.793) 1.115 (0.723–1.721) HR hazard ratio 1 3 Int J Clin Oncol (2017) 22:1042–1049 1047 1.0 evaluated the utility of the GPS in unresectable LAESCC 0.9 because these markers can be easily analyzed using typical 0.8 blood tests. The GPS, which takes into account the serum Absent 0.7 Present 0.6 CRP and albumin levels, simply reflects the systemic HR =1.63(95%CI 1.12 –2.36) (Univariable analysis) 0.5 inflammation status of patients with cancer and their prog- 0.4 nosis, as described for various cancer types. In addition, a 0.3 0.2 series of work by McMillan and other groups outlined the 0.1 indisputable association between an increased CRP level 0.0 01234567 8 and poor survival in various tumor types [10, 11, 24–29]. Years after randomization No. at Risk The present study shows that the pretreatment GPS was an Absent 70 48 28 22 16 12 6 2 0 independent prognostic factor in patients with unresect- Present 61 29 15 8 4 4 3 2 0 able LAESCC. These findings suggest the pivotal role of systemic inflammation in patients with EC who undergo CRT for LAESCC. Most reports regarding GPS for ESCC Fig. 3 Kaplan–Meier estimates of overall survival in the absence of stenosis (n = 70) and in the presence of stenosis (n = 61) investigated patients who underwent esophagectomy and described the GPS as reflecting deeper tumor involvement, higher numbers of nodal metastases, and more advanced 1.0 cancer stage [13, 30, 31]. These findings suggest that the 0.9 0.8 GPS of patients with unresectable LAESCC might repre- GPS0 0.7 GPS1 sent a synthetic marker for the prediction of overall sur- 0.6 GPS2 vival because it reflects both tumor factors and patient 0.5 HR = 1.22 (95%CI 0.80 –1.86) for GPS 1 vs GPS 0 0.4 HR = 1.95(95%CI 1.19 –3.18) for GPS 2 vs GPS 0 background. Recently, in a retrospective study at a single 0.3 institution, Ohira et  al. reported that GPS1/2 scores were 0.2 0.1 closely associated with poor prognosis compared with 0.0 patients with a GPS0 score with the same cancer stage as 01234567 8 Years after randomization No. at Risk those in our study [32]. GPS 0 56 37 20 17 11 8 5 2 0 GPS 1 48 30 17 10 8 7 4 2 0 This analysis does have some limitations. One of the lim- 10 63 1 1 00 GPS 2 27 0 itations is that the patients enrolled in this clinical trial, who might have a better prognosis, are compared with patients in clinical practice. Therefore, other factors that affect the Fig. 4 Kaplan–Meier estimates of overall survival in GPS0 (n = 56), prognosis of “real” patients may be obscured. Another GPS1 (n = 48) and GPS2 (n = 27) patients limitation is the definition of “stenosis”. In the JCOG0303 trial, stenosis of the primary lesion was not strictly defined, chemoradiotherapy [18], locoregional control [19, 20], and and the clinical investigators who participated in this study survival [17, 19, 21, 22] after radiation therapy for EC with diagnosed it subjectively. Thus, “stenosis” seemed to vary or without chemotherapy. In the present study, the pretreat- among investigators. Some investigators reported stenosis ment Hb level did not demonstrate clinical significance in in patients who experienced dysphagia, while other inves- terms of its association with survival, which may have been tigators reported stenosis in patients whose primary lesions because included patients were limited to those who had were too narrow for an endoscopy to be performed. Col- met the eligibility criteria of the prospective study. lecting and evaluating data in an objective manner using Thus far, SCC and CEA are well-known tumor mark- the dysphagia score (DS) would be a better procedure to ers for esophageal cancer [23]. Nevertheless, their signifi- follow in the next clinical trial [33]. cance as prognostic factors is controversial. In this study, In conclusion, esophageal stenosis was thought to be using univariable and multivariable analyses, the relation- a potential stratification factor in a randomized trial that ship between survival and the levels of CEA and SCC was included patients with unresectable LAESCC. Further- examined; however, they were not found to be prognostic more, the GPS represents a prognostic factor in patients factors. with LAESCC who are treated with CRT. These data can Patients with advanced EC comprise a heterogeneous be easily collected before treatment. Further studies are population, as represented by differences in variables such needed to clarify the association between these pretreat- as age, co-morbidities, primary tumor location, and ECOG ment data and clinical outcomes and to be able to use these PS. Additionally, simple and inexpensive markers are parameters as stratification factors in the next prospective desirable to establish a prognosis of ESCC. Therefore, we randomized trial of CRT for LAESCC. 1 3 OverallSurvival OverallSurvival 1048 Int J Clin Oncol (2017) 22:1042–1049 Acknowledgements This study was supported by a National Can- 11. Brown DJF, Milroy R, Preston T et  al (2007) The relationship cer Center Research and Development Fund (26-A-4) from the Min- between an inflammation-based prognostic score (Glasgow istry of Health, Labour and Welfare of Japan and the Daininaika- Prognostic Score) and changes in serum biochemical variables Doumonkai (grant to T.O.). We thank Junki Mizusawa for providing in patients with advanced lung and gastrointestinal cancer. J Clin invaluable statistical support. We also thank Dr. Kenichi Miyamoto Pathol 60:705–708. doi:10.1136/jcp.2005.033217 and Hisato Kawakami for reviewing the manuscript. 12. Ishizuka M, Nagata H, Takagi K et al (2007) Inflammation-based prognostic score is a novel predictor of postoperative outcome Compliance with ethical standards in patients with colorectal cancer. Ann Surg 246:1047–1051. doi:10.1097/SLA.0b013e3181454171 13. Kobayashi T, Teruya M, Kishiki T et  al (2008) Inflammation- Conflict of interest Dr. Haruhiko Fukuda has received honoraria based prognostic score, prior to neoadjuvant chemoradiother- from Taiho Pharma, Co., Ltd and Chugai Pharmaceutical Co., Ltd. apy, predicts postoperative outcome in patients with esophageal squamous cell carcinoma. Surgery 144:729–735. doi:10.1016/j. Open Access This article is distributed under the terms of the Crea- surg.2008.08.015 tive Commons Attribution 4.0 International License (http://crea- 14. Roxburgh CS, Crozier JE, Maxwell F et  al (2009) Comparison tivecommons.org/licenses/by/4.0/), which permits unrestricted use, of tumour-based (Petersen index) and inflammation-based (Glas- distribution, and reproduction in any medium, provided you give gow Prognostic Score) scoring systems in patients undergoing appropriate credit to the original author(s) and the source, provide a curative resection for colon cancer. Br J Cancer 100:701–706. link to the Creative Commons license, and indicate if changes were doi:10.1038/sj.bjc.6604926 made. 15. Sharma R, Hook J, Kumar M et  al (2008) Evaluation of an inflammation-based prognostic score in patients with advanced ovarian cancer. Eur J Cancer 44:251–256. doi:10.1016/j. ejca.2007.11.011 References 16. Sobin LH, Fleming ID (1997) International Union Against Can- cer TNM classification of malignant tumours, 5th edn. Springer, Berlin 1. Enzinger PC, Mayer RJ (2003) Esophageal cancer. N Engl J Med 17. Zenda S, Hironaka S, Boku N et  al (2008) Impact of hemo- 349:2241–2252. doi:10.1056/NEJMra035010 globin level on survival in definitive chemoradiotherapy for T4/ 2. Ozawa S, Tachimori Y, Baba H et al (2010) Comprehensive reg- M1 lymph node esophageal cancer. Dis Esophagus 21:195–200. istry of esophageal cancer in Japan, 2002. Esophagus 7:7–22. doi:10.1111/j.1442-2050.2007.00750.x doi:10.1007/s10388-010-0228-6 18. Yi Y, Li B, Sun H et al (2010) Predictors of sensitivity to chemo- 3. Matsubara T, Ueda M, Kokudo N et  al (2001) Role of radiotherapy of esophageal squamous cell carcinoma. Tumour esophagectomy in treatment of esophageal carcinoma with clini- Biol 31:333–340. doi:10.1007/s13277-010-0041-9 cal evidence of adjacent organ invasion. World J Surg 25:279– 19. Zhao KL, Liu G, Jiang GL et  al (2006) Association of haemo- 284. doi:10.1007/s002680020060 globin level with morbidity and mortality of patients with locally 4. Ikeda K, Ishida K, Sato N et  al (2001) Chemoradiotherapy fol- advanced oesophageal carcinoma undergoing radiotherapy—a lowed by surgery forthoracic esophageal cancer potentially or secondary analysis of three consecutive clinical phase III trials. actually involving adjacent organs. Dis Esophagus 14:197–201. Clin Oncol 18:621–627. doi:10.1016/j.clon.2006.04.015 doi:10.1046/j.1442-2050.2001.00184.x 20. Rades D, Golke H, Schild SE et al (2008) The impact of tumor 5. Ishida K, Ando N, Yamamoto S et  al (2004) Phase II study of expression of erythropoietin receptors and erythropoietin on cisplatin and 5-fluorouracil with concurrent radiotherapy in clinical outcome of esophageal cancer patients treated with advanced squamous cell carcinoma of the esophagus: a Japan chemoradiation. Int J Radiat Oncol Biol Phys 71:152–159. Esophageal Oncology Group (JEOG)/Japan Clinical Oncol- doi:10.1016/j.ijrobp.2007.09.027 ogy Group Trial (JCOG9516). Jpn J Clin Oncol 34:615–619. 21. Zhang HQ, Wang RB, Yan HJ et  al (2012) Prognostic signifi- doi:10.1093/jjco/hyh107 cance of CYFRA21-1, CEA and hemoglobin in patients with 6. Fujita H, Sueyoshi S, Tanaka T et  al (2005) Esophagectomy: is esophageal squamous cancer undergoing concurrent chemora- it necessary after chemoradiotherapy for a locally advanced T4 diotherapy. Asian Pac J Cancer Prev 13:199–203. doi:10.7314/ esophageal cancer? Prospective nonrandomized trial comparing APJCP.2012.13.1.199 chemoradiotherapy with surgery versus without surgery. World J 22. Valencia Julve J, Alonso Orduña V, Escó Barón R et  al (2006) Surg 29:25–30. doi:10.1007/s00268-004-7590-2 Influence of hemoglobin levels on survival after radical treat- 7. Ohtsu A, Boku N, Muro K et  al (1999) Definitive chemoradio- ment of esophageal carcinoma with radiotherapy. Clin Transl therapy for T4 and/or M1 lymph node squamous cell carcinoma Oncol 8:22–30. doi:10.1007/s12094-006-0091-z of the esophagus. J Clin Oncol 17:2915–2921 23. Zhang J, Zhu Z, Liu Y et  al (2015) Diagnostic value of multi- 8. Ishikura S, Ohtsu A, Shirao K et  al (2005) A phase I/II study ple tumor markers for patients with esophageal carcinoma. PLoS of nedaplatin and 5-fluorouracil with concurrent radiotherapy One 10:e0116951. doi:10.1371/journal.pone.0116951 in patients with T4 esophageal cancer: Japan Clinical Oncology 24. Forrest LM, McMillan DC, McArdle CS et al (2004) Compari- Group Trial (JCOG 9908). Esophagus 2:133–137. doi:10.1007/ son of an inflammation-based prognostic score (GPS) with per - s10388-005-0055-3 formance status (ECOG) in patients receiving platinum-based 9. Shinoda M, Ando N, Kato K et al (2015) Randomized study of chemotherapy for inoperable non-small-cell lung cancer. Br J low-dose versus standard-dose chemoradiotherapy for unresect- Cancer 90:1704–1706. doi:10.1038/sj.bjc.6601789 able esophageal squamous cell carcinoma (JCOG0303). Cancer 25. Ramsey S, Lamb GW, Aitchison M et al (2007) Evaluation of an Sci 106:407–412. doi:10.1111/cas.12622 inflammation-based prognostic score in patients with metastatic 10. McMillan DC (2009) Systemic inflammation, nutri- renal cancer. Cancer 109:205–212. doi:10.1002/cncr.22400 tional status and survival in patients with cancer. Curr 26. McMillan DC, Crozier JEM, Canna K et  al (2007) Evaluation Opin Clin Nutr Metab Care 12:223–226. doi:10.1097/ of an inflammation-based prognostic score (GPS) in patients MCO.0b013e32832a7902 1 3 Int J Clin Oncol (2017) 22:1042–1049 1049 undergoing resection for colon and rectal cancer. Int J Colorectal squamous cell carcinoma. Surg Today 44:875–883. doi:10.1007/ Dis 22:881–886. doi:10.1007/s00384-006-0259-6 s00595-013-0628-z 27. Glen P, Jamieson NB, McMillan DC et  al (2006) Evalua- 31. Feng JF, Zhao Q, Chen QX (2014) Prognostic significance of tion of an inflammation-based prognostic score in patients Glasgow prognostic score in patients undergoing esophagectomy with inoperable pancreatic cancer. Pancreatology 6:450–453. for esophageal squamous cell carcinoma. Saudi J Gastroenterol doi:10.1159/000094562 20:48–53. doi:10.4103/1319-3767.126319 28. Crumley AB, McMillan DC, McKernan M et al (2006) Evalua- 32. Ohira M, Kubo N, Masuda G et  al (2015) Glasgow prognostic tion of an inflammation-based prognostic score in patients with score as a prognostic clinical marker in T4 esophageal squamous inoperable gastro-oesophageal cancer. Br J Cancer 94:637–641. cell carcinoma. Anticancer Res 35:4897–4901 doi:10.1038/sj.bjc.6602998 33. Ogilvie AL, Dronfield MW, Ferguson R et  al (1982) Palliative 29. Al Murri AM, Bartlett JM, Canney PA et al (2006) Evaluation of intubation of oesophagogastric neoplasms at fibreoptic endos- an inflammation-based prognostic score (GPS) in patients with copy. Gut 23:1060–1067. doi:10.1136/gut.23.12.1060 metastatic breast cancer. Br J Cancer 9:227–230 30. Nakamura M, Iwahashi M, Nakamori M et  al (2014) New prognostic score for the survival of patients with esophageal 1 3 http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png International Journal of Clinical Oncology Pubmed Central

Esophageal stenosis and the Glasgow Prognostic Score as independent factors of poor prognosis for patients with locally advanced unresectable esophageal cancer treated with chemoradiotherapy (exploratory analysis of JCOG0303)

Loading next page...
 
/lp/pubmed-central/esophageal-stenosis-and-the-glasgow-prognostic-score-as-independent-FGfAGWYaiv

References (36)

Publisher
Pubmed Central
Copyright
© The Author(s) 2017
ISSN
1341-9625
eISSN
1437-7772
DOI
10.1007/s10147-017-1154-6
Publisher site
See Article on Publisher Site

Abstract

Int J Clin Oncol (2017) 22:1042–1049 DOI 10.1007/s10147-017-1154-6 ORIGINAL ARTICLE Esophageal stenosis and the Glasgow Prognostic Score as independent factors of poor prognosis for patients with locally advanced unresectable esophageal cancer treated with chemoradiotherapy (exploratory analysis of JCOG0303) 1 2 3 4 2 Tatsuya Okuno  · Masashi Wakabayashi  · Ken Kato  · Masayuki Shinoda  · Hiroshi Katayama  · 5 6 7 8 9 10 Hiroyasu Igaki  · Yasuhiro Tsubosa  · Takashi Kojima  · Hiroshi Okabe  · Yusuke Kimura  · Tatsuyuki Kawano  · 11 12 5 2 2 13 Shinichi Kosugi  · Yasushi Toh  · Hoichi Kato  · Kenichi Nakamura  · Haruhiko Fukuda  · Satoshi Ishikura  · 14 15 Nobutoshi Ando  · Yuko Kitagawa  · Japan Esophageal Oncology Group/Japan Clinical Oncology Group Received: 28 December 2016 / Accepted: 14 June 2017 / Published online: 17 July 2017 © The Author(s) 2017. This article is an open access publication Abstract CRP ≤1.0  mg/dL and albumin ≥3.5  g/dL were classified Background The aim of this study was to investigate the as GPS0. If only CRP was increased or only albumin was possible prognostic factors and predictive accuracy of the decreased, the patients were classified as GPS1, and the Glasgow Prognostic Score (GPS) for patients with unre- patients with CRP >1.0  mg/dL and albumin <3.5  g/dL sectable locally advanced esophageal squamous cell carci- were classified as GPS2. noma (LAESCC) treated with chemoradiotherapy. Results The patients’ backgrounds were as follows: median Methods One hundred forty-two patients were enrolled age (range), 62 (37–75); male/female, 119/12; ECOG PS in JCOG0303 and assigned to the standard cisplatin and 0/1/2, 64/65/2; and clinical stage (UICC 5th) IIB/III/IVA/ 5-fluorouracil (PF)-radiotherapy (RT) group or the low- IVB, 3/75/22/31. Multivariable analyses indicated only dose PF-RT group. One hundred thirty-one patients with esophageal stenosis as a common factor for poor prognosis. sufficient data were included in this analysis. A Cox regres - In addition, overall survival tended to decrease according to sion model was used to analyze the prognostic factors of the GPS subgroups (median survival time (months): GPS0/ patients with unresectable LAESCC treated with PF-RT. GPS1/GPS2 16.1/14.9/8.7). The GPS was classified based on the baseline C-reactive Conclusions Esophageal stenosis was identified as a protein (CRP) and serum albumin levels. Patients with candidate stratification factor for randomized trials of * Tatsuya Okuno Department of Surgery, Kyoto University Graduate School t-okuno@pg7.so-net.ne.jp of Medicine, Kyoto, Japan Department of Surgery, Iwate Medical University School Division of Gastroenterology, Department of Internal of Medicine, Morioka, Japan Medicine, Graduate School of Medicine, Kobe University, 7-5-1 Kusunoki Chuo, Kobe 650-0017, Japan Tokyo Medical and Dental University, Tokyo, Japan Japan Clinical Oncology Group Data Center/Operations Department of Surgery, Niigata University Medical Office, National Cancer Center, Tokyo, Japan and Dental Hospital, Niigata, Japan 3 12 Gastrointestinal Medical Oncology Division, National Department of Gastroenterological Surgery, National Kyushu Cancer Center Hospital, Tokyo, Japan Cancer Center, Fukuoka, Japan 4 13 Department of Gastrointestinal Surgery, Aichi Cancer Center Department of Radiology, Koshigaya Municipal Hospital, Hospital, Nagoya, Japan Koshigaya, Saitama, Japan 5 14 Esophageal Surgery Division, National Cancer Center Department of Surgery, International Goodwill Hospital, Hospital, Tokyo, Japan Yokohama, Japan 6 15 Division of Esophageal Surgery, Shizuoka Cancer Center Department of Surgery, Keio University, School of Medicine, Hospital, Shizuoka, Japan Tokyo, Japan Gastrointestinal Oncology Division, National Cancer Center Hospital East, Kashiwa, Japan Vol:.(1234567890) 1 3 Int J Clin Oncol (2017) 22:1042–1049 1043 unresectable LAESCC patients. Furthermore, GPS repre- in therapeutic effectiveness and prognosis is important sents a prognostic factor for LAESCC patients treated with for tailoring an optimized treatment strategy and further chemoradiotherapy. improving treatment outcomes. Emerging evidence sug- Clinical Trial Information UMIN000000861. gests that TNM stage, a weight loss of more than 10% of one’s body mass, dysphagia, large tumors, older age, Keywords T4 esophageal cancer · Chemoradiotherapy · and lymphatic micrometastases (identified by immuno- Prognostic factor · Esophageal stenosis · Glasgow histochemical analysis) are independent predictors of a Prognostic Score poor prognosis in patients with advanced EC who have undergone surgery [1]. However, little is known about the prognostic factors in patients with advanced unresectable Introduction LAESCC who have been treated with CRT. Therefore, we investigated clinical pretreatment factors that might affect Despite substantial improvements in screening and diag- the survival of patients enrolled in the JCOG0303 trial. nosis, esophageal cancer (EC) is frequently diagnosed at Moreover, hypoalbuminemia and C-reactive protein a very advanced stage [1]. The thoracic esophagus lacks (CRP) are associated with a poor prognosis in cancer serosa and is closely surrounded by the trachea, bronchus, patients [10]. The Glasgow Prognostic Score (GPS) com- lung and aorta. Thus, esophageal cancer is more likely bines albumin and CRP into a risk stratification score to invade these vital organs and become unresectable. for the prognosis of clinical outcome in cancer patients. According to the Comprehensive Registry of Esophageal This scoring system has been validated in colorectal can- Cancer in Japan, the incidence of cT4 esophageal carci- cer (CRC) and other malignant tumors, including EC noma is approximately 15% among all EC patients [2]. [11–15]. Given that hypoalbuminemia and CRP are asso- Curative resection is not feasible in patients with locally ciated with poor prognosis in cancer patients [10], we advanced esophageal squamous cell carcinoma (LAESCC) also evaluated the potential prognostic role of the GPS on if the cancer has invaded other organs, and such cases long-term outcome in patients undergoing definitive CRT have an unfavorable prognosis [3–6]. Definitive chemora- for LAESCC. diotherapy (CRT) is the standard of care that is currently available for unresectable locally advanced EC [5–8]. In a former phase II trial, 18 of 54 (33%) patients with clini- cal T4 and/or M1 lymph nodes (M1Lym) and esophageal Methods squamous cell carcinoma (ESCC), who received cisplatin and 5-fluorouracil (PF) with concurrent 60-Gy irradia- Schema of the JCOG0303 Study tion, achieved a complete response (CR); in that study, the median overall survival (OS) and the 3-year survival rate Key eligibility criteria were as follows: age of 75  years were 9  months and 23%, respectively [7]. Additionally, a or younger, patients with histologically proven squamous new multicenter trial JCOG0303 was conducted to evalu- cell carcinoma, adenosquamous, or basaloid carcinoma ate the efficacy and toxicity, and particularly the long-term of the thoracic esophagus, and an Eastern Cooperative outcome, of CRT in patients with unresectable LAESCC. Oncology Group (ECOG) performance status (PS) of The aim of JCOG0303 was to evaluate whether RT plus 0–2; in addition, patients who had any of the following low-dose cisplatin and 5-fluorouracil (LDPF) chemother - conditions were included: definite clinical T4 cancer, at apy had an advantage in terms of survival and/or toxicity least 1 unresectable metastatic regional lymph node due over RT plus standard-dose cisplatin and 5-fluorouracil to invasion into an adjacent organ, or computed tomog- (SDPF) chemotherapy in a randomized phase II/III trial. raphy (CT)-based evidence of M1 Lym disease, such as In a primary analysis of the phase II portion, the study fixed supraclavicular or celiac lymph nodes. Regional was terminated due to futility. In an updated analysis, the lymph nodes were defined using criteria specified by the median OS and the 3-year OS rate were 13.1 months and 5th edition of the Union for International Cancer Control 25.9%, respectively, in the SDPF-RT arm, and 14.4 months (UICC) TNM staging system [16]. After confirmation and 25.7%, respectively, in the LDPF-RT arm. The pri- of eligibility, the patients were randomized at the JCOG mary endpoint of OS was nearly equivalent in both treat- Data Center. Written informed consent was obtained ment arms [9]. from all enrolled patients, and the institutional review The status of each patient with invasive thoracic esoph- boards of all participating institutions approved the study ageal cancer varies widely due to variability in the extent protocol. The JCOG0303 trial was registered with the of the tumors and the nutritional and general status of the UMIN Clinic Trials Registry (http://www.umin.ac.jp/ctr/) patients. Thus, establishing the factors that are involved with the identification number UMIN000000861. 1 3 1044 Int J Clin Oncol (2017) 22:1042–1049 Treatment patients whose CRP level was >1.0 mg/dL and whose albu- min level was <3.5 g/dL were classified as GPS2. The patients were randomized to receive either SDPF-RT (arm A) or LDPF-RT (arm B). The chemotherapy regimen in arm A consisted of 70  mg⁄m cisplatin given on days 1 Results and 29 combined with a continuous infusion of 700 mg⁄m 5-FU given on days 1–4 and 29–32. Patients in arm B Patient characteristics received a 1  h infusion of 4  mg⁄m cisplatin before RT, combined with a continuous infusion of 200  mg⁄m 5-FU In all, 142 patients from 24 participating institutions of on the first 5 days of each week. Treatment completion was the Japan Esophageal Oncology Group of the JCOG were defined as the termination of two courses of chemotherapy enrolled between May 2000 and May 2006. Of these, 71 and 60 Gy of radiotherapy within 63 days. patients were randomly assigned to SDPF-RT (arm A) and 71 patients were randomly assigned to LDPF-RT (arm B). Statistical analysis Of the patients randomized into the two groups, 7 in arm A and 4 in arm B were excluded from this ancillary study For the analysis of prognostic factors using pretreatment because there were no documented data for the aforemen- factors, the OS was calculated from the date of randomiza- tioned prognostic factors. Thus, 131 patients comprised tion until death from any cause, or the OS was censored at the population for the analysis based on prognostic fac- the time of the last follow-up. An initial univariable Cox tors (Fig.  1). The patient characteristics in each group that regression analysis was performed to evaluate factors that was subjected to the analysis using pretreatment factors are could potentially affect the survival outcomes. Further - shown in Table 1. more, a multivariable analysis that included all covariates The median overall survival times (MSTs) of all rand- and a multivariable analysis with the stepwise selection omized 142 patients and of the final 131 patients included method were performed. in this study were 14.4  months. Among all randomized The following pretreatment factors were tested in the patients, the 1- and 3-year survival rates were 57.0 and analyses: age (≥65 vs. <65  years), sex (male vs. female), 23.9%, respectively, while those rates of the patients ECOG PS (0 vs. 1 or 2), clinical T stage (T1, T2, T3 vs. included in this study were 58.8 and 22.9%, respectively T4), clinical N stage (N0 vs. N1), clinical M stage (M0 vs. (Fig. 2). M1a or M1b), histological subtype of SCC (well-differen- tiated or moderately differentiated vs. poorly differentiated vs. unknown), location of the primary tumor (upper vs. All registered patients middle or lower), clinically diagnosed esophageal steno- (n = 142) sis (absent vs. present), adjacent organ invasion mediated by LN metastasis (absent vs. present), intramural metasta- SDPF-RT: 71 sis (absent vs. present), and the laboratory data described LDPF-RT: 71 below, which were obtained upon enrollment in the trial. The white blood cell count, neutrophil count, hemoglobin level (Hb), CRP level, albumin level, GPT, BUN, cre- Exclusion of 11 patients atinine clearance (CrC), CEA level, and SCC level were dichotomized, with cutoff points at 10,000/mm , 8000/ with missing data forat mm , 13  g/dL, 1.0  mg/dL, 3.5  g/dL, 30  IU/L, 15  mg/dL, least 1 baseline factor 80 mL/min, 5 ng/mL, and 1.5 ng/mL, respectively. The cut- Phase II portion (n = 2) off points of each laboratory test were determined based on a clinical perspective. Survival curves were estimated using the Kaplan–Meier method. Two-sided p values <0.05 were All analyzed patients considered statistically significant. All statistical analyses (n = 131) were performed with SAS9.2 (SAS Institute, Cary, NC, USA). SDPF-RT: 64 The GPS was determined as previously described by LDPF-RT: 67 several groups [12–15]. Patients with a CRP level ≤1.0 mg/ dL and an albumin level ≥3.5 g/dL were classified as GPS0. If only the CRP level was increased or if only the albumin level was decreased, patients were classified as GPS1, and Fig. 1 Patient flowchart 1 3 Int J Clin Oncol (2017) 22:1042–1049 1045 Table 1 Baseline patient characteristics (n = 131) histological subtype of SCC (unknown) were significantly associated with poor survival. A multivariable analysis Characteristic SDPF-RT (n = 64) LDPF-RT (n = 67) with stepwise selection also confirmed esophageal stenosis Gender as an indicator of poor survival, although a serum Alb level  Male 61 58 <3.5 g/dL was not an independent prognostic factor in the  Female 3 9 multivariable analysis that included all covariates. The haz- Age (years) ard ratio for patients with esophageal stenosis (present) was  Median 63 62 1.60 (95% CI 1.10–2.32) compared with patients without  Range 37–75 43–74 esophageal stenosis (absent) (p = 0.014) in the multivaria- ECOG PS ble analysis with stepwise selection (Table 2). Patients with  0 30 34 esophageal stenosis had a median survival of 9.9  months,  1 33 32 while patients without esophageal stenosis had a median  2 1 1 survival of 16.0 months (p = 0.001) (Fig. 3). cTNM (UICC 5th edition)  T1 2 0 GPS as an independent prognostic factor for survival  T2 5 2  T3 9 12 The association between survival and the GPS in patients  T4 48 53 with esophageal carcinoma is shown in Fig.  4. The HR  N0 9 9 for overall survival in GPS1 patients to GPS0 patients  N1 55 58 was 1.22 (95% CI 0.80–1.86; p  =  0.35) and was 1.95  M0 38 40 (95% CI 1.19–3.18; p  =  0.008) in GPS2 patients to GPS0  M1a 8 14 patients. Patients with a GPS of 0 had a median survival  M1b 18 13 of 16.1 months, while patients with a GPS of 1 and 2 had Clinical stage (UICC) a median survival of 14.9 and 8.7  months, respectively  IIb 2 1 (p = 0.001) (Fig. 4). These findings suggest that an elevated  III 36 39 GPS is associated with poor prognosis. Moreover, we con-  IVa 8 14 firmed that the presence or absence of stenosis and the GPS  IVb 18 13 score did not have remarkable relationship. ECOG PS Eastern Cooperative Oncology Group performance status Discussion 1.0 Here, we retrospectively investigated the therapeutic out- 0.9 comes of patients who were prospectively registered in the 0.8 JCOG0303 trial and evaluated pretreatment clinical param- 0.7 All randomized 0.6 Pts included eters that might consistently affect survival. In the present 0.5 HR = 1.08 (95%CI 0.55 –2.13) study, two multivariable analyses revealed esophageal 0.4 0.3 stenosis as the only single factor that was associated with 0.2 poor prognosis in patients with unresectable LAESCC. To 0.1 our knowledge, this is the first study to report an associa- 0.0 01234567 8 tion between treatment outcome and clinically diagnosed Years after randomization No. at Risk All randomized 142 81 47 34 23 18 11 5 0 esophageal stenosis. Pts included 131 77 43 30 20 16 9 4 0 Currently, the established factors of a poor prognosis are PS, tumor size and TNM stage, but in our analysis, no correlations were found between these factors and prog- Fig. 2 Kaplan–Meier estimates of the overall survival of all rand- nosis. This might reflect the fact that our patients in the omized (n = 142) patients, including the study (n = 131) patients present study were limited to those with T4 and/or unre- sectable supraclavicular lymph node metastasis (i.e., those Multivariable analysis at a relatively advanced stage with a poor prognosis). Zenda et  al. described that low pretreatment levels of Hb A multivariable analysis that included all covariates dem- were significantly associated with the efficacy of CRT in onstrated that esophageal stenosis, lower Hb, clinical M1 patients with T4/M1 EC [17]. Similarly, low levels of Hb stage, location of the primary tumor (middle or lower) and have been reported to be associated with sensitivity to 1 3 OverallSurvival 1046 Int J Clin Oncol (2017) 22:1042–1049 Table 2 Multivariable Cox regression analysis of overall survival according to pretreatment factors Factors Levels N Univariable analysis Multivariable analysis Including all covariates Stepwise selection HR (95% CI) p value HR (95% CI) p value HR (95% CI) p value Age 65> 85 1 0.320 1 0.231 65≤ 46 1.219 (0.825–1.802) 1.335 (0.832–2.140) Sex Female 12 1 0.464 1 0.744 Male 119 0.784 (0.409–1.504) 0.881 (0.414–1.878) PS 0 64 1 0.446 1 0.360 1 or 2 67 1.156 (0.796–1.678) 1.228 (0.791–1.908) T-stage T4 101 1 0.558 1 0.177 T1, T2, T3 30 0.873 (0.555–1.373) 0.646 (0.343–1.217) N-stage N0 18 1 0.348 1 0.976 N1 113 1.308 (0.747–2.292) 1.011 (0.512–1.996) M-stage M0 78 1 0.070 1 0.002 M1a or M1b 53 1.426 (0.972–2.092) 2.271 (1.368–3.772) Location of the pri- Ut 37 1 0.396 1 0.009 mary tumor Mt or Lt 94 1.199 (0.788–1.825) 2.068 (1.198–3.569) Histological subtype Well or Moderately 97 1 All (0.372) 1 All (0.024) of SCC differentiated Poorly differentiated 19 1.138 (0.680–1.903) 0.623 1.311 (0.741–2.320) 0.352 Unknown 15 1.494 (0.844–2.644) 0.168 2.844 (1.332–6.070) 0.007 Stenosis Absent 70 1 0.010 1 0.002 1 0.014 Present 61 1.629 (1.123–2.361) 2.014 (1.293–3.137) 1.598 (1.101–2.319) Adjacent organ Inva- Absent 105 1 0.822 1 0.545 sion mediated by LN Present 26 1.056 (0.657–1.698) 1.193 (0.675–2.108) metastasis Intramural metastasis Absent 107 1 All (0.081) 1 All (0.216) Present 13 1.641 (0.913–2.948) 0.098 1.883 (0.927–3.824) 0.080 Unknown 11 1.735 (0.922–3.266) 0.088 1.057 (0.456–2.451) 0.897 WBC 10,000≥ 113 1 0.611 1 0.095 10000< 18 0.868 (0.503–1.498) 0.449 (0.176–1.148) ANC 8000≥ 121 1 0.920 1 0.533 8000< 10 1.036 (0.520–2.065) 1.477 (0.434–5.027) Hb 13≥ 85 1 0.042 1 0.008 13< 46 0.664 (0.447–0.986) 0.509 (0.310–0.836) CRP 1.0 (mg/dL)≥ 69 1 0.117 1 0.367 1.0 (mg/dL)< 62 1.344 (0.929–1.945) 1.266 (0.758–2.114) Alb 3.5 (g/dL)> 40 1 0.018 1 0.988 1 0.025 3.5 (g/dL)≤ 91 0.621 (0.420–0.921) 0.996 (0.559–1.773) 0.636(0.429–0.944) GPT 30> 105 1 0.032 1 0.134 30≤ 26 1.639 (1.044–2.573) 1.531 (0.878–2.670) BUN 15.0> 98 1 0.863 1 0.336 15.0≤ 33 0.963 (0.628–1.478) 1.269 (0.781–2.063) CrC 80.0> 65 1 0.405 1 0.077 80.0≤ 66 1.171 (0.807–1.698) 1.500 (0.957–2.352) CEA 5> 110 1 0.849 1 0.733 5≤ 21 0.954 (0.587–1.550) 0.902 (0.501–1.626) SCC 1.5≥ 67 1 0.265 1 0.622 1.5< 64 1.236 (0.852–1.793) 1.115 (0.723–1.721) HR hazard ratio 1 3 Int J Clin Oncol (2017) 22:1042–1049 1047 1.0 evaluated the utility of the GPS in unresectable LAESCC 0.9 because these markers can be easily analyzed using typical 0.8 blood tests. The GPS, which takes into account the serum Absent 0.7 Present 0.6 CRP and albumin levels, simply reflects the systemic HR =1.63(95%CI 1.12 –2.36) (Univariable analysis) 0.5 inflammation status of patients with cancer and their prog- 0.4 nosis, as described for various cancer types. In addition, a 0.3 0.2 series of work by McMillan and other groups outlined the 0.1 indisputable association between an increased CRP level 0.0 01234567 8 and poor survival in various tumor types [10, 11, 24–29]. Years after randomization No. at Risk The present study shows that the pretreatment GPS was an Absent 70 48 28 22 16 12 6 2 0 independent prognostic factor in patients with unresect- Present 61 29 15 8 4 4 3 2 0 able LAESCC. These findings suggest the pivotal role of systemic inflammation in patients with EC who undergo CRT for LAESCC. Most reports regarding GPS for ESCC Fig. 3 Kaplan–Meier estimates of overall survival in the absence of stenosis (n = 70) and in the presence of stenosis (n = 61) investigated patients who underwent esophagectomy and described the GPS as reflecting deeper tumor involvement, higher numbers of nodal metastases, and more advanced 1.0 cancer stage [13, 30, 31]. These findings suggest that the 0.9 0.8 GPS of patients with unresectable LAESCC might repre- GPS0 0.7 GPS1 sent a synthetic marker for the prediction of overall sur- 0.6 GPS2 vival because it reflects both tumor factors and patient 0.5 HR = 1.22 (95%CI 0.80 –1.86) for GPS 1 vs GPS 0 0.4 HR = 1.95(95%CI 1.19 –3.18) for GPS 2 vs GPS 0 background. Recently, in a retrospective study at a single 0.3 institution, Ohira et  al. reported that GPS1/2 scores were 0.2 0.1 closely associated with poor prognosis compared with 0.0 patients with a GPS0 score with the same cancer stage as 01234567 8 Years after randomization No. at Risk those in our study [32]. GPS 0 56 37 20 17 11 8 5 2 0 GPS 1 48 30 17 10 8 7 4 2 0 This analysis does have some limitations. One of the lim- 10 63 1 1 00 GPS 2 27 0 itations is that the patients enrolled in this clinical trial, who might have a better prognosis, are compared with patients in clinical practice. Therefore, other factors that affect the Fig. 4 Kaplan–Meier estimates of overall survival in GPS0 (n = 56), prognosis of “real” patients may be obscured. Another GPS1 (n = 48) and GPS2 (n = 27) patients limitation is the definition of “stenosis”. In the JCOG0303 trial, stenosis of the primary lesion was not strictly defined, chemoradiotherapy [18], locoregional control [19, 20], and and the clinical investigators who participated in this study survival [17, 19, 21, 22] after radiation therapy for EC with diagnosed it subjectively. Thus, “stenosis” seemed to vary or without chemotherapy. In the present study, the pretreat- among investigators. Some investigators reported stenosis ment Hb level did not demonstrate clinical significance in in patients who experienced dysphagia, while other inves- terms of its association with survival, which may have been tigators reported stenosis in patients whose primary lesions because included patients were limited to those who had were too narrow for an endoscopy to be performed. Col- met the eligibility criteria of the prospective study. lecting and evaluating data in an objective manner using Thus far, SCC and CEA are well-known tumor mark- the dysphagia score (DS) would be a better procedure to ers for esophageal cancer [23]. Nevertheless, their signifi- follow in the next clinical trial [33]. cance as prognostic factors is controversial. In this study, In conclusion, esophageal stenosis was thought to be using univariable and multivariable analyses, the relation- a potential stratification factor in a randomized trial that ship between survival and the levels of CEA and SCC was included patients with unresectable LAESCC. Further- examined; however, they were not found to be prognostic more, the GPS represents a prognostic factor in patients factors. with LAESCC who are treated with CRT. These data can Patients with advanced EC comprise a heterogeneous be easily collected before treatment. Further studies are population, as represented by differences in variables such needed to clarify the association between these pretreat- as age, co-morbidities, primary tumor location, and ECOG ment data and clinical outcomes and to be able to use these PS. Additionally, simple and inexpensive markers are parameters as stratification factors in the next prospective desirable to establish a prognosis of ESCC. Therefore, we randomized trial of CRT for LAESCC. 1 3 OverallSurvival OverallSurvival 1048 Int J Clin Oncol (2017) 22:1042–1049 Acknowledgements This study was supported by a National Can- 11. Brown DJF, Milroy R, Preston T et  al (2007) The relationship cer Center Research and Development Fund (26-A-4) from the Min- between an inflammation-based prognostic score (Glasgow istry of Health, Labour and Welfare of Japan and the Daininaika- Prognostic Score) and changes in serum biochemical variables Doumonkai (grant to T.O.). We thank Junki Mizusawa for providing in patients with advanced lung and gastrointestinal cancer. J Clin invaluable statistical support. We also thank Dr. Kenichi Miyamoto Pathol 60:705–708. doi:10.1136/jcp.2005.033217 and Hisato Kawakami for reviewing the manuscript. 12. Ishizuka M, Nagata H, Takagi K et al (2007) Inflammation-based prognostic score is a novel predictor of postoperative outcome Compliance with ethical standards in patients with colorectal cancer. Ann Surg 246:1047–1051. doi:10.1097/SLA.0b013e3181454171 13. Kobayashi T, Teruya M, Kishiki T et  al (2008) Inflammation- Conflict of interest Dr. Haruhiko Fukuda has received honoraria based prognostic score, prior to neoadjuvant chemoradiother- from Taiho Pharma, Co., Ltd and Chugai Pharmaceutical Co., Ltd. apy, predicts postoperative outcome in patients with esophageal squamous cell carcinoma. Surgery 144:729–735. doi:10.1016/j. Open Access This article is distributed under the terms of the Crea- surg.2008.08.015 tive Commons Attribution 4.0 International License (http://crea- 14. Roxburgh CS, Crozier JE, Maxwell F et  al (2009) Comparison tivecommons.org/licenses/by/4.0/), which permits unrestricted use, of tumour-based (Petersen index) and inflammation-based (Glas- distribution, and reproduction in any medium, provided you give gow Prognostic Score) scoring systems in patients undergoing appropriate credit to the original author(s) and the source, provide a curative resection for colon cancer. Br J Cancer 100:701–706. link to the Creative Commons license, and indicate if changes were doi:10.1038/sj.bjc.6604926 made. 15. Sharma R, Hook J, Kumar M et  al (2008) Evaluation of an inflammation-based prognostic score in patients with advanced ovarian cancer. Eur J Cancer 44:251–256. doi:10.1016/j. ejca.2007.11.011 References 16. Sobin LH, Fleming ID (1997) International Union Against Can- cer TNM classification of malignant tumours, 5th edn. Springer, Berlin 1. Enzinger PC, Mayer RJ (2003) Esophageal cancer. N Engl J Med 17. Zenda S, Hironaka S, Boku N et  al (2008) Impact of hemo- 349:2241–2252. doi:10.1056/NEJMra035010 globin level on survival in definitive chemoradiotherapy for T4/ 2. Ozawa S, Tachimori Y, Baba H et al (2010) Comprehensive reg- M1 lymph node esophageal cancer. Dis Esophagus 21:195–200. istry of esophageal cancer in Japan, 2002. Esophagus 7:7–22. doi:10.1111/j.1442-2050.2007.00750.x doi:10.1007/s10388-010-0228-6 18. Yi Y, Li B, Sun H et al (2010) Predictors of sensitivity to chemo- 3. Matsubara T, Ueda M, Kokudo N et  al (2001) Role of radiotherapy of esophageal squamous cell carcinoma. Tumour esophagectomy in treatment of esophageal carcinoma with clini- Biol 31:333–340. doi:10.1007/s13277-010-0041-9 cal evidence of adjacent organ invasion. World J Surg 25:279– 19. Zhao KL, Liu G, Jiang GL et  al (2006) Association of haemo- 284. doi:10.1007/s002680020060 globin level with morbidity and mortality of patients with locally 4. Ikeda K, Ishida K, Sato N et  al (2001) Chemoradiotherapy fol- advanced oesophageal carcinoma undergoing radiotherapy—a lowed by surgery forthoracic esophageal cancer potentially or secondary analysis of three consecutive clinical phase III trials. actually involving adjacent organs. Dis Esophagus 14:197–201. Clin Oncol 18:621–627. doi:10.1016/j.clon.2006.04.015 doi:10.1046/j.1442-2050.2001.00184.x 20. Rades D, Golke H, Schild SE et al (2008) The impact of tumor 5. Ishida K, Ando N, Yamamoto S et  al (2004) Phase II study of expression of erythropoietin receptors and erythropoietin on cisplatin and 5-fluorouracil with concurrent radiotherapy in clinical outcome of esophageal cancer patients treated with advanced squamous cell carcinoma of the esophagus: a Japan chemoradiation. Int J Radiat Oncol Biol Phys 71:152–159. Esophageal Oncology Group (JEOG)/Japan Clinical Oncol- doi:10.1016/j.ijrobp.2007.09.027 ogy Group Trial (JCOG9516). Jpn J Clin Oncol 34:615–619. 21. Zhang HQ, Wang RB, Yan HJ et  al (2012) Prognostic signifi- doi:10.1093/jjco/hyh107 cance of CYFRA21-1, CEA and hemoglobin in patients with 6. Fujita H, Sueyoshi S, Tanaka T et  al (2005) Esophagectomy: is esophageal squamous cancer undergoing concurrent chemora- it necessary after chemoradiotherapy for a locally advanced T4 diotherapy. Asian Pac J Cancer Prev 13:199–203. doi:10.7314/ esophageal cancer? Prospective nonrandomized trial comparing APJCP.2012.13.1.199 chemoradiotherapy with surgery versus without surgery. World J 22. Valencia Julve J, Alonso Orduña V, Escó Barón R et  al (2006) Surg 29:25–30. doi:10.1007/s00268-004-7590-2 Influence of hemoglobin levels on survival after radical treat- 7. Ohtsu A, Boku N, Muro K et  al (1999) Definitive chemoradio- ment of esophageal carcinoma with radiotherapy. Clin Transl therapy for T4 and/or M1 lymph node squamous cell carcinoma Oncol 8:22–30. doi:10.1007/s12094-006-0091-z of the esophagus. J Clin Oncol 17:2915–2921 23. Zhang J, Zhu Z, Liu Y et  al (2015) Diagnostic value of multi- 8. Ishikura S, Ohtsu A, Shirao K et  al (2005) A phase I/II study ple tumor markers for patients with esophageal carcinoma. PLoS of nedaplatin and 5-fluorouracil with concurrent radiotherapy One 10:e0116951. doi:10.1371/journal.pone.0116951 in patients with T4 esophageal cancer: Japan Clinical Oncology 24. Forrest LM, McMillan DC, McArdle CS et al (2004) Compari- Group Trial (JCOG 9908). Esophagus 2:133–137. doi:10.1007/ son of an inflammation-based prognostic score (GPS) with per - s10388-005-0055-3 formance status (ECOG) in patients receiving platinum-based 9. Shinoda M, Ando N, Kato K et al (2015) Randomized study of chemotherapy for inoperable non-small-cell lung cancer. Br J low-dose versus standard-dose chemoradiotherapy for unresect- Cancer 90:1704–1706. doi:10.1038/sj.bjc.6601789 able esophageal squamous cell carcinoma (JCOG0303). Cancer 25. Ramsey S, Lamb GW, Aitchison M et al (2007) Evaluation of an Sci 106:407–412. doi:10.1111/cas.12622 inflammation-based prognostic score in patients with metastatic 10. McMillan DC (2009) Systemic inflammation, nutri- renal cancer. Cancer 109:205–212. doi:10.1002/cncr.22400 tional status and survival in patients with cancer. Curr 26. McMillan DC, Crozier JEM, Canna K et  al (2007) Evaluation Opin Clin Nutr Metab Care 12:223–226. doi:10.1097/ of an inflammation-based prognostic score (GPS) in patients MCO.0b013e32832a7902 1 3 Int J Clin Oncol (2017) 22:1042–1049 1049 undergoing resection for colon and rectal cancer. Int J Colorectal squamous cell carcinoma. Surg Today 44:875–883. doi:10.1007/ Dis 22:881–886. doi:10.1007/s00384-006-0259-6 s00595-013-0628-z 27. Glen P, Jamieson NB, McMillan DC et  al (2006) Evalua- 31. Feng JF, Zhao Q, Chen QX (2014) Prognostic significance of tion of an inflammation-based prognostic score in patients Glasgow prognostic score in patients undergoing esophagectomy with inoperable pancreatic cancer. Pancreatology 6:450–453. for esophageal squamous cell carcinoma. Saudi J Gastroenterol doi:10.1159/000094562 20:48–53. doi:10.4103/1319-3767.126319 28. Crumley AB, McMillan DC, McKernan M et al (2006) Evalua- 32. Ohira M, Kubo N, Masuda G et  al (2015) Glasgow prognostic tion of an inflammation-based prognostic score in patients with score as a prognostic clinical marker in T4 esophageal squamous inoperable gastro-oesophageal cancer. Br J Cancer 94:637–641. cell carcinoma. Anticancer Res 35:4897–4901 doi:10.1038/sj.bjc.6602998 33. Ogilvie AL, Dronfield MW, Ferguson R et  al (1982) Palliative 29. Al Murri AM, Bartlett JM, Canney PA et al (2006) Evaluation of intubation of oesophagogastric neoplasms at fibreoptic endos- an inflammation-based prognostic score (GPS) in patients with copy. Gut 23:1060–1067. doi:10.1136/gut.23.12.1060 metastatic breast cancer. Br J Cancer 9:227–230 30. Nakamura M, Iwahashi M, Nakamori M et  al (2014) New prognostic score for the survival of patients with esophageal 1 3

Journal

International Journal of Clinical OncologyPubmed Central

Published: Jul 17, 2017

There are no references for this article.