Access the full text.
Sign up today, get DeepDyve free for 14 days.
Loading next page...
/lp/american-journal-of-neuroradiology/post-covid-19-brain-18f-fdg-pet-findings-a-retrospective-single-center-pGA0hzIY30
References
References for this paper are not available at this time. We will be adding them shortly, thank you for your patience.
- Publisher
- American Journal of Neuroradiology
- Copyright
- © 2023 by American Journal of Neuroradiology. Indicates open access to non-subscribers at www.ajnr.org
- ISSN
- 0195-6108
- eISSN
- 1936-959X
- DOI
- 10.3174/ajnr.a7863
- Publisher site
- See Article on Publisher Site
Abstract
Post-COVID-19 Brain [18F] FDG-PET Findings: A Retrospective Single-Center Study in the United States P. Debs, N. Khalili, L. Solnes, A. Al-Zaghal, H.I. Sair, V. Yedavalli and L.P. Luna AJNR Am J Neuroradiol 2023, 44 (5) 517-522 This information is current as doi: https://doi.org/10.3174/ajnr.A7863 of June 3, 2023. http://www.ajnr.org/content/44/5/517 ORIGINAL RESEARCH ADULT BRAIN Post-COVID-19 Brain [ F] FDG-PET Findings: A Retrospective Single-Center Study in the United States P. Debs, N. Khalili, L. Solnes, A. Al-Zaghal, H.I. Sair, V. Yedavalli, and L.P. Luna ABSTRACT BACKGROUND AND PURPOSE: The pathophysiology of neurologic manifestations of postacute sequelae of Severe Acute Respiratory Syndrome coronavirus 2 (SARS-CoV-2) infection is not clearly understood. Our aim was to investigate brain metabolic activity on [ F] FDG-PET/CT scans in patients with a history of coronavirus disease 2019 (COVID-19) infection before imaging. MATERIALS AND METHODS: This retrospective study included 45 patients who underwent [ F] FDG-PET/CT imaging for any reason and had, at least once, tested positive for COVID-19 at any time before imaging. Fifteen patients had available [ F] FDG-PET scans obtained under identical conditions before the infection. A group of 52 patients with melanoma or multiple myeloma who underwent [ F] FDG-PET/CT were used as controls. Whole-brain 2-sample t test analysis was performed using SPM software to identify clusters of hypo- and hypermetabolism and compare brain metabolic activity between patients with COVID-19 and controls. Paired sample t test comparison was also performed for 15 patients, and correlations between metabolic values of clusters and clinical data were measured. RESULTS: Compared with the control group, patients with a history of COVID-19 infection exhibited focal areas of hypometabolism in the bilateral frontal, parietal, occipital, and posterior temporal lobes and cerebellum (P ¼ .05 uncorrected at the voxel level, family-wise error–corrected at the cluster level) that peaked during the first 2 months, improved to near-complete recovery around 6 months, and disappeared at 12 months. Hypermetabolism involving the brainstem, cerebellum, limbic structures, frontal cortex, and periventricular white matter was observed only at 2-6 months after infection. Older age, neurologic symptoms, and worse disease severity scores positively correlated with the metabolic changes. CONCLUSIONS: This study demonstrates a profile of time-dependent brain PET hypo- and hypermetabolism in patients with con- firmed SARS-CoV-2 infection. ABBREVIATIONS: BMI ¼ body mass index; COVID-19 ¼ coronavirus disease 2019; FWE ¼ family-wise error; GLM ¼ general linear model; neuro-PASC ¼ neurologic manifestations of postacute sequelae of SARS-CoV-2 infection; PASC ¼ postacute sequelae of SARS-CoV-2; PCR ¼ polymerase chain reaction; RT-PCR ¼ reverse transcriptase PCR; SARS-CoV-2 ¼ Severe Acute Respiratory Syndrome coronavirus 2; Tmax ¼ time-to-maximum s of December 2022, Severe Acute Respiratory Syndrome co- ageusia, and anosmia; nevertheless, the reality of the long-term 3-5 Aronavirus 2 (SARS-CoV-2) has resulted in .97 million con- consequences of COVID-19 is becoming more evident, and firmed cases in the United States, with approximately 1 million many survivors of COVID-19 experience chronic postviral 3,6 all-time deaths, the highest among all countries and the six- complications. 1,2 teenth-highest per 100,000 population worldwide. Knowledge Evidence of persistent neurologic symptoms following acute of coronavirus disease 2019 (COVID-19) mainly focuses on the COVID-19 is increasing, and this process was recently termed acute illness and related symptoms such as cough, fever, myalgia, neurologic manifestations of postacute sequelae of SARS-CoV-2 3,7 infection (neuro-PASC). The most common neurologic mani- Received January 27, 2023; accepted after revision March 5. festations are “brain fog”, fatigue, headache, numbness/tingling, 3,8,9 From the Russell H. Morgan Department of Radiology and Radiological Science, dysgeusia, and anosmia. Although neuro-PASC is more fre- Johns Hopkins University School of Medicine, Baltimore, Maryland. 3,10,11 quent in patients needing hospitalization, studies have Please address correspondence to Licia P. Luna, MD, PhD, Russell H. Morgan Department of Radiology and Radiological Science, Johns Hopkins Hospital, shown that postacute sequelae of SARS-CoV-2 (PASC) can also Division of Neuroradiology, 600 N Wolfe St, Phipps B100F, Baltimore, MD, 21287; impact children, young adults, and those who experience only e-mail: lluna6@jhmi.edu; @liciapluna mild COVID-19 symptoms and do not require respiratory sup- Indicates open access to non-subscribers at www.ajnr.org 3,7,12 port or hospitalization. The mechanism behind SARS-CoV-2-induced pathologic Indicates article with online supplemental data. 13,14 http://dx.doi.org/10.3174/ajnr.A7863 changes in the CNS is still unclear. Two main hypotheses AJNR Am J Neuroradiol 44:517–22 May 2023 www.ajnr.org 517 attempt to explain this mechanism: indirect effects via peripheral Clinical and laboratory data were extracted from the patients’ inflammation and direct effects via SARS-CoV-2 CNS invasion. electronic medical records. Variables at the initial infection were On the one hand, a cytokine storm (ie, an inflated immune age, sex, medical comorbidities, and presenting symptoms. When response instigated by the infection) might play an indirect role patients required hospital admission, information regarding sup- in the neurologic manifestations of PASC; on the other hand, plemental oxygenation, admission to the intensive care unit, and some reports suggest that SARS-CoV-2 may directly invade the the development of new symptoms was also collected. COVID-19 CNS and possibly infect brain cells via the functional receptor cases were rated as mild, moderate, or severe according to interna- human angiotensin-converting enzyme 2. tional guidelines and using the National Early Warning Score 2. Despite these 2 hypotheses, questions regarding the mecha- Variables at the time of imaging included the time lag between nisms underlying the pathophysiology of neuro-PASC symptoms the first COVID-19 test and the PET/CT scan, any persistent remain unanswered, and imaging could help elucidate the under- neurologic symptoms, body mass index (BMI), glucose levels, lying processes. PET might meaningfully contribute to our medications, administered [ F] FDG dose, and administration of understanding of the pathophysiologic changes in patients post- any COVID-19 vaccine before imaging. Neurologic symptoms COVID-19 by identifying affected brain regions. The underly- included persistent headaches, memory impairment, difficulty ing mechanisms can be explored by investigating changes in met- concentrating, fatigue, and insomnia. The presence of neurologic abolic parameters, and [ F] FDG-PET/CT can be a valuable tool symptoms at the time of imaging was defined as persistent symp- for detecting or ruling out severe coexistent processes and high- toms following COVID-19 infection that were reported in clinical lighting alterations in brain metabolism. notes dating no more than 7 days before PET imaging. Prior studies assessing imaging patterns associated with post- COVID-19 functional symptoms in PET/CT scans have com- Patients with Pre-COVID Imaging. Among included patients, 15 pared whole-brain voxel-based analysis with a local database of had available [ F] FDG-PET scans obtained under identical con- healthy individuals to characterize cerebral metabolism pat- ditions before the COVID-19 infection. Information on BMI, glu- 17-20 terns. However, to our knowledge, no prior studies have cose levels, medications, and administered [ F] FDG pertaining assessed the potential alterations in brain metabolism in patients to the pre-COVID scan dates was also collected for these patients. without functional symptoms at the time of imaging. We present a retrospective analysis of brain [ F] FDG-PET scans of patients Controls. A group of 52 age- and sex-matched controls who with a biologically confirmed SARS-CoV-2 infection with and underwent PET/CT imaging for initial staging and were scanned without persistent functional symptoms. On the basis of our clin- under identical conditions before January 2021 was selected from ical observations, we hypothesized that a history of COVID-19 our institutional database as follows: those with recently diagnosed infection may be associated with functional brain involvement melanoma or multiple myeloma, oncologically negative brain that can be identified using [ F] FDG-PET. We aimed to com- images, no history of neurologic disorders, and no recent use of pare PET scans using whole-brain voxel-based analysis with a local database of patients imaged under similar circumstances to psychotropic medications. All scans were obtained before adminis- characterize cerebral metabolism patterns, assess temporal evolu- tering any systemic chemotherapy or anticancer drugs. The limit tion, and correlate PET abnormalities with patient characteristics date (December 31, 2020) used to select controls was settled on the and functional symptoms. basis of the latest epidemiologic data on the pandemic in the United States to avoid any potential bias. Information on controls regarding age, sex, medical comorbidities, medications, BMI, glu- MATERIALS AND METHODS cose levels, and administered [ F] FDG dose was subsequently Patient Selection Post-COVID Patients. The institutional review board approved collected from the electronic medical records. this retrospective single-center study and waived the requirement for written informed consent. The study complied with the [ F] FDG-PET Imaging and Processing [ F] FDG-PET scans were acquired in the same center using an Declaration of Helsinki and Health Insurance Portability and acquisition protocol conforming to guidelines put forth by the Accountability Act (HIPPA) regulations. American College of Radiology, the American College of Nuclear Forty-five patients underwent whole-body [ F] FDG-PET Medicine, and the Society of Nuclear Medicine and Molecular imaging between April 2020 and October 2021 at our institution. Imaging. Whole-body scans were acquired on a Biograph mCT Inclusion criteria were adult patients older than 18 years of age (Siemens; 36 patients and 47 controls) or a Discovery DRX or with at least 1 documented case of SARS-CoV-2 infection con- DLS (GE Healthcare; 9 patients and 5 controls) with in-line CT firmed by polymerase chain reaction (PCR) at any time before for attenuation correction at 60 minutes after the [ F] FDG in- imaging. None of the patients had an active SARS-CoV-2 infec- tion at time of imaging. We excluded patients with incomplete travenous administration in individuals fasting for at least 4 PET/CT imaging of the entire brain, patients with brain imaging hours with a controlled, normal glycemic level. Brain images showing major structural abnormalities (eg, tumors, prior sur- were extracted from the whole-body [ F] FDG-PET/CT as previ- 20,23,24 gery, ischemic infarcts, cerebral venous thrombosis) unrelated to ously described and converted to Analyze format for pre- the COVID-19 infection, and patients with documented neuro- processing in SPM software (http://www.fil.ion.ucl.ac.uk/spm/ logic/psychiatric antecedents or symptoms preceding the SARS- software/spm12). The Montreal Neurological Institute 152 brain template was used as the standard template for registering CoV-2 infection (ie, confounding clinical variables). The study flow diagram is shown in the Online Supplemental Data. PET/CT images to T1-weighted MR images using the FMRIB 518 Debs May 2023 www.ajnr.org Subjects’ baseline characteristics Distinct t-contrasts identified brain Post-COVID Controls areas where glucose metabolism was (n = 45) (n = 52) significantly lower or higher in the Age (mean) (range) (yr) 58 (18–87) 57 (24–73) patient group than in the control group. Male sex (No.) (%) 24 (53.33) 28 (53.85) Patients post-COVID were further sub- [ F] FDG administered dose (mean) (SD) (mCi) 12.28 (3.18) 11.88 (2.59) 2 grouped according to the time delay BMI (mean) (SD) (kg/m ) 28.62 (7.59) 28.45 (6.36) between the positive reverse transcrip- Glucose levels (mean) (SD) (mg/dL) 95.13 (22.90) 98.04 (23.82) Diabetes (No.) (%) 3 (6.67) 6 (11.54) tase PCR (RT-PCR) and PET imaging: High blood pressure (No.) (%) 14 (31.11) 21 (40.38) 0–2, 2–6, 6–12, and.12 months. Delay between positive RT-PCR and imaging (mean) 6.57 (4.85) – Correlations between metabolic val- (SD) (mo) uesof clustersand demographic and clin- COVID-19 severity mild/moderate or severe (No.) (%) 37/8 (82.22/17.78) – ical data were measured for group-level Hospitalization (No.) (%) 15 (33.33%) Oxygen supplementation (No.) (%) 9 (20.00%) comparisons. A general linear model Mechanical ventilation (No.) (%) 2 (4.44%) (GLM) of the preprocessed FDG-PET COVID-19 vaccine at any time before [ F] FDG-PET 28 (62.2) – data of patients taken as 1 group was scan (No.) (%) constructed. Age, sex, COVID-19 sever- Symptoms (No.) (%) ity, presence of neurologic symptoms (eg, Dysosmia/dysgeusia 17 (37.80) – Fever 16 (35.6) – syncope, loss of taste and smell, general- Chills 6 (13.3) – ized weakness, difficulty concentrating) Cough 16 (35.6) – at the time of infection and at the time of Dyspnea 8 (17.8) – imaging and the time lag between imag- Chest pain 2 (4.4) – ing and infection were introduced as Pharyngitis 4 (8.9) – Rhinitis 3 (6.7) – covariates of interest centered around Headache 4 (8.9) – condition means in separate GLMs. Fatigue 7 (15.6) – Separate t-contrasts identified brain Muscular pain 1 (2.2) – regions showing significant positive or Dysgeusia 1 (2.2) – negative correlations between the covari- Anosmia/hyposmia 1 (2.2) – Diarrhea 4 (8.9) – ates of interest and regional cerebral glu- Loss of appetite 2 (4.4) – cose metabolism. The SPM t-statistic Neurologic symptoms at time of imaging (No.) (%) (SPM[T]) maps were acquired at an un- Persistent headaches 3 (6.7) corrected height threshold (voxel-level Memory impairment 1 (2.2) significance) of P, .05, with a correction Difficulty concentrating 1 (2.2) Fatigue 1 (2.2) for multiple comparisons at the level of Insomnia 1 (2.2) the cluster using the family-wise error Note:—The en dash indicates not applicable. (FWE) rate for a corrected P value, .05. No statistically significant difference was observed between patients post-COVID and sex- and age-matched controls (P , .05, 2-sample t test for continuous variables, x test for dichotomous variables). Between the time of the initial diagnosis and PET imaging. All other symptoms pertain to the time of the initial RESULTS diagnosis. Clinical Characteristics of Patients Persistent symptoms following COVID-19 infection that were reported in the patients’ records dating no more than 7 days before PET imaging. Post-COVID The clinical characteristics of the 45 included post-COVID patients are Linear Image Registration Tool (FLIRT; http://www.fmrib.ox.ac. detailed in the Table. No significant difference was found 26,27 uk/fsl/fslwiki/FLIRT). Nonlinear spatial normalization of between the post-COVID and control groups in pre-PET glucose images to a specific [ F] FDG-PET template in the Montreal levels (mean, 95.13 [SD, 22.90] versus 98.04 [SD, 23.82]); mean Neurological Institute space was performed using the SPM8 soft- BMI (mean, 28.62 [SD, 7.59] versus 28.45 [SD, 6.36]); and admin- ware. The images were then smoothed with a Gaussian kernel istered [ F] FDG activity (mean, 12.28 [SD, 3.18] versus 11.88 with a full width at half maximum of 8 8 8 mm to increase [SD, 2.59] mCi). The mean time delay between positive RT-PCR 28,29 results and imaging was 6.57 (SD, 4.85) months (range, 1–24 the SNR. Global activity normalization was performed by months). At the time of diagnosis, the median National Early proportional scaling (Online Supplemental Data). All images Warning 2 clinical score was 2, corresponding to a low-risk grad- were checked for the presence of nonperfect fits before analysis. ing; 33.3% of patients required hospitalization (15/45), and 4.4% required mechanical ventilation (2/45). Single-Subject and Group-Level [ F] FDG-PET SPM Analyses Whole-brain 2-sample t test analysis was initially performed at Among the post-COVID patients, fever and cough were the the voxel level using SPM8 software to compare patients with most common manifestations at presentation (16/45, 35.6% controls and identify clusters of hypometabolism in the patient each). Dyspnea, fatigue, and chills were reported in 8, 7, and 6 group. Paired sample t test (single-subject) comparison between patients, respectively (17.8%, 15.6%, and 13.3%). On admission, 1 patients before and after COVID-19 infection was also performed. patient had of loss of taste (1/45, 2.2%), while another had loss of AJNR Am J Neuroradiol 44:517–22 May 2023 www.ajnr.org 519 smell (1/45, 2.2%). A total of 17 patients reported experiencing Correlations between Brain Hypometabolism and Clinical loss of smell and taste between the initial diagnosis of COVID-19 Variables and their PET imaging (17/45, 37.8%). At the time of imaging, a Age, neurologic symptoms at the time of imaging, and SARS- total of 6 patients (13%) reported $1 persistent neurologic symp- CoV-2 infection-severity scores were all significantly associated tom following COVID-19 infection, including persistent head- with the widespread extension and severity of brain metabolic aches (3/45, 6.7%), memory issues, focus impairment, fatigue, patterns seen on post-COVID versus control scans. Precisely, and insomnia (1/45 each, 2.2%) (Table). older age, the presence of neurologic symptoms, and worse dis- ease severity scores were positively correlated with the degree of hypometabolism in the bilateral parietal, posterior frontal, and Patterns of [ F] FDG-PET Hypo- and Hypermetabolism All groups showed comparable global metabolism uptake values temporal lobes, as well as the degree of hypermetabolism in the (Online Supplemental Data). Specifically, the comparison central cerebral and subcortical regions (Online Supplemental between the whole post-COVID (n ¼ 45), pre-COVID (n ¼ 15), Data). In addition, SARS-CoV-2 infection severity scores were and control patient (n ¼ 52) groups showed no significant differ- also significantly associated with the widespread extension and severity of brain metabolic patterns seen on post- versus pre- ences (P ¼ .18). COVID scans (Online Supplemental Data). PET Metabolic Profile of Post-COVID Scans Compared with Controls DISCUSSION Clusters of significant hypo- and hypermetabolism in the whole This whole-brain voxel-based PET study demonstrates brain group of post-COVID patients and subgroup analyses are reported metabolic abnormalities in patients with a history of biologically in the Online Supplemental Data. In comparison with the 52 con- confirmed SARS-CoV-2 infection. Voxel-based brain analysis trols, post-COVID patients presented with significant hypometabo- supported the recent hypothesis on SARS-CoV-2 infection- lism (P-voxel, .05 uncorrected, P-cluster, .05, FWE-corrected) related brain metabolic impairment and provided new insights involving the bilateral parietal lobes, including the precuneus into the pathophysiology of COVID-related brain abnormalities. regions (time-to maximum [Tmax] ¼ 3.90); frontal lobes, includ- Cross-sectional subgroup analyses suggest that the brain metabo- ing the anterior cingulate (Tmax ¼ 2.40) and prefrontal cortices lism remains mildly altered 6 months after disease onset and (Tmax ¼ 3.04); occipital lobes (Tmax ¼ 3.90); right temporal lobe gradually improves after 6–12 months. Moreover, the significant (Tmax ¼ 2.64); and right cerebellum (Tmax ¼ 2.74). The whole brain hypermetabolism observed between 2 and 6 months after group of post-COVID patients did not exhibit brain regions of infection in subcortical brain regions, including the limbic struc- statistically significant relative hypermetabolism compared with tures (eg, hippocampi and amygdala), in patients post-COVID controls. In addition, cross-sectional imaging subgroup analyses compared with both the control group and the pre-COVID imag- showed more severe and extensive brain hypometabolism during ing data set suggests that brain inflammation peaks and subse- the first 2 months after the infection onset, followed by a progressive quently recovers after this time window. return to normal metabolic activity. At 6–12 months, patients A gradual change in the severity and extent of hypometabo- showed a near-complete recovery of brain abnormalities, with resid- lism from a widespread pattern in ,2 months to a limited ual limited hypometabolic clusters in the anterior cingulate cortex, involvement of the anterior cingulate cortex, right orbitofrontal posterior inferior frontal gyri, right frontal operculum, and right cortex, bilateral posterior gyrus rectus, right insula, and medial temporal-insular region. The significantly reduced metabolism dis- temporal lobes was observed between 6 and 12 months postin- appeared at 12 months. Hypermetabolism involving the brainstem, fection. However, no significant hypometabolism was noted in cerebellum, limbic structures (ie, amygdala and hippocampus bilat- these particular regions in the 0- to 2- and 2- to 6-month sub- erally), a smaller region of the frontal cortex, and periventricular groups, raising the possibility that these regions might be prefer- white matter was observed only at 2–6 months after infection entially impaired in the later stages of the disease. Conversely, (Online Supplemental Data). significant areas of hypermetabolism involving the brainstem, cerebellum, and limbic structures were observed in the 2- to 6- PET Metabolic Profile of Patients Post- versus Pre-COVID month period subgroup. One explanation for this finding is that At the group level, SPM paired samples t test comparison the regions of hypometabolism involving the frontoinsular areas between PET scans of patients pre- and post-COVID (n ¼ 15, and limbic system after 6 months may result from a peak in the P-voxel,.05 uncorrected, P-cluster,.05 FWE-corrected) did active inflammation involving the central regions of the brain not show statistically significant clusters. However, at 2–6months between 2 and 6 months after disease onset. Correspondingly, after the infection onset, post-COVID patients presented with a prior studies have reported a predominant involvement of the significant decrease in regional glucose consumption in the bilat- brainstem, cerebellum, and limbic structures (amygdala and 17,20,30 eral parietal lobes, posterior frontal lobes including the frontal hippocampus bilaterally) as hypermetabolic hallmarks. eye fields and the left cingulate cortex, and occipital lobes com- Most interesting, this metabolic pattern was observed in our pared with their pre-COVID scans (n ¼ 6). Moreover, significant nonselected whole-patient group regardless of the individual’s hypermetabolic areas were found in the bilateral limbic structures neurologic symptoms or lack thereof, highlighting some degree (anterior hippocampi and amygdala), brainstem, ventral thalami, of abnormal cerebral metabolic activity possibly occurring even left inferior frontal lobe, and left cerebellum during the 2- to 6- in the absence of clinical symptoms or during the subclinical month interval (Online Supplemental Data). phase of the infection. Yet, because previous prospective studies 520 Debs May 2023 www.ajnr.org had focused on comparing clinically symptomatic patients with direct and indirect effects on brain metabolism and could, there- healthy controls without including asymptomatic post-COVID fore, affect the validity of our results. Third, the inclusion of patients, it remains unclear to which degree clinical symptoms patients who have had at least 1 dose of the COVID-19 vaccine 17-20 correlate with specific abnormal brain metabolic patterns. before PET/CT may have impacted the clinical disease course To the best of our knowledge, no regional brain metabolic and clinical-imaging correlations. Fourth, the 15 patients with studies assessing intrasubject variability in patients before and after pre-COVID imaging could have potentially had another earlier COVID-19 infections have been previously published. Therefore, undocumented COVID-19 infection before the pre-COVID we believe this approach reduces the potential bias of pre-existing imaging, thus affecting the validity of that comparison. regional metabolic changes in these individuals. Even though no significant regional metabolic changes were observed in the CONCLUSIONS whole-group level analysis at the statistical threshold used, signifi- 18 The present study demonstrates a pattern of reversible brain PET cant regional [ F] FDG-PET hypometabolism and hypermetabo- hypo- and hypermetabolic changes in patients with confirmed lism in the post-COVID subgroup of patients (n ¼ 6) imaged SARS-CoV-2 infection. The degree of observed alterations between 2 and 6 months after COVID-19 infection onset were appears to be transient and positively correlates with older age, noted. The lack of significant findings on [ F] FDG-PET in the neurologic symptoms at the time of imaging, and worse disease whole patient group and other subgroups might be partly due to severity scores. Even so, metabolic changes were detected in the FWE at the cluster level used and the small sample size, asymptomatic patients and patients without symptoms strongly because the FWE rate may be considered conservative for group- suggestive of COVID-19. Brain imaging could potentially serve level analyses with small sample sizes. as a tool to identify asymptomatic cases whenever incidental [ F] Correlation analyses revealed a relationship between regional FDG-PET/CT findings suspicious for SARS-CoV-2 are detected cerebral glucose metabolism and age, neurologic symptoms at the and to further understand the pathophysiology behind the neuro- time of imaging, and SARS-CoV-2 infection severity. This is con- logic symptoms observed with long COVID. Future studies must sistent with previous studies documenting a correlation between clarify how PET imaging can be used as a potential biomarker to the severity of SARS-CoV-2-related loss of smell and metabolic follow-up patients clinically, monitor disease progression, and changes and an inverse correlation between an increased number assess recovery. of functional symptoms with alterations in the brainstem and cer- 19,31 ebellum metabolism. In addition, increased age has been Disclosure forms provided by the authors are available with the full text and reported as a predictor of more severe outcomes, explaining the PDF of this article at www.ajnr.org. correlation between age and the observed metabolic derange- 32,33 ments. Moreover, neurologic symptoms at the time of imag- REFERENCES ing also correlated with brain metabolic changes, in line with a 1. Johns Hopkins Coronavirus Resource Center. Mortality Analyses. study by Goehringer et al showing correlations between metab- March 10, 2023. https://coronavirus.jhu.edu/data/mortality.Accessed olism in the orbitofrontal, internal temporal area, and the pons December 5, 2022 ROIs and multiple cognitive assessment scales. Also, the presence 2. WHO Coronavirus (COVID-19). 2022. https://covid19.who.int/table. Accessed December 5, 2022 of persistent neurologic symptoms such as headaches, memory 3. Rudroff T, Workman CD, Ponto LLB. 18F-FDG-PET imaging for and concentration impairment, fatigue, and insomnia was posi- post-COVID-19 brain and skeletal muscle alterations. Viruses tively correlated with decreased metabolism in the parietal lobes. 2021;13:2283 CrossRef Medline As previously assessed, the strong correlation between changes in 4. Havervall S, Rosell A,Phillipson M,et al. Symptoms and functional cerebral metabolism, specifically in the frontoparietal regions, and impairment assessed 8 months after mild COVID-19 among health the cognitive functions served by those regions in COVID-19 ence- care workers. JAMA 2021;325:2015–16 CrossRef Medline 30,35 phalopathy suggests that these changes are interdependent. 5. Maltezou HC, Raftopoulos V, Vorou R, et al. Association between Limitations must be considered for the interpretation of this upper respiratory tract viral load, comorbidities, disease severity, and outcome of patients with SARS-CoV-2 infection. JInfect Dis study. First, this retrospective, observational study evaluated a 2021;223:1132–38 CrossRef Medline small group of patients who underwent FDG-PET imaging for 6. Leung TY, Chan AY, Chan EW, et al. Short-and potential long-term other clinical indications, most of which were oncologic. Patients adverse health outcomes of COVID-19: a rapid review. Emerg with brain lesions who underwent chemotherapy or radiation Microbes Infect 2020;9:2190–99 CrossRef Medline therapy to the head and neck area were excluded to reduce the 7. Carfì A, Bernabei R, Landi F; Gemelli Against COVID-19 Post-Acute confounding effect of such comorbidities. Second, patients Care Study Group. Persistent symptoms in patients after acute selected for the control group were diagnosed with either multi- COVID-19. JAMA 2020;324:603–05 CrossRef Medline ple myeloma or melanoma, which could have affected the study 8. Goërtz YM, Herck MV, Delbressine JM, et al. Persistent symptoms 3 months after a SARS-CoV-2 infection: the post-COVID-19 syn- results. Enrolling disease-free patients was not possible due to the drome? ERJ Open Res 2020;6:00542-2020 CrossRef Medline retrospective nature of the study, because patients usually 9. Graham EL, Clark JR, Orban ZS, et al. Persistent neurologic symp- undergo PET imaging for a clinical indication. To minimize bias, toms and cognitive dysfunction in non-hospitalized Covid-19 “long we selected patients who had recently been diagnosed, were not haulers.” AnnClinTranslNeurol 2021;8:1073–85 CrossRef Medline experiencing symptoms, had not commenced any treatment, and 10. Mehta P, McAuley DF, Brown M, et al; HLH Across Speciality did not show evidence of brain abnormalities on imaging. Collaboration, UK. COVID-19: consider cytokine storm syndromes Nonetheless, both melanoma and multiple myeloma can have and immunosuppression. Lancet 2020;395:1033–34 CrossRef Medline AJNR Am J Neuroradiol 44:517–22 May 2023 www.ajnr.org 521 11. Song E, Zhang C, Israelow B, et al. Neuroinvasion of SARS-CoV-2 in 23. Hua C, Merchant TE, Li X, et al. Establishing age-associated norma- human and mouse brain. JExp Med 2021;218:e20202135 CrossRef tive ranges of the cerebral 18F-FDG uptake ratio in children. J Medline Nucl Med 2015;56:575–79 CrossRef Medline 24. Huang YC, Hsu CC, Huang P, et al. The changes in brain metabo- 12. Mandal S, Barnett J, Brill SE, et al; ARC Study Group. ‘Long- lism in people with activated brown adipose tissue: a PET study. COVID’: a cross-sectional study of persisting symptoms, bio- Neuroimage 2011;54:142–47 CrossRef Medline marker and imaging abnormalities following hospitalisation for 25. Friston KJ, Holmes AP, Worsley KJ, et al. Statistical parametric COVID-19. Thorax 2021;76:396–98 CrossRef Medline maps in functional imaging: a general linear approach. Hum Brain 13. Paterson RW, Brown RL, Benjamin L, et al. The emerging spectrum Mapp 1994;2:189–210 CrossRef of COVID-19 neurology: clinical, radiological and laboratory find- 26. Jenkinson M, Bannister P, Brady M, et al. Improved optimization ings. Brain 2020;143:3104–20 CrossRef Medline for the robust and accurate linear registration and motion correc- 14. Rubin R. As their numbers grow, COVID-19 “long haulers” stump tion of brain images. Neuroimage 2002;17:825–41 CrossRef Medline experts. JAMA 2020;324:1381–83 CrossRef Medline 27. Jenkinson M, Smith S. A global optimisation method for robust 15. Fontana IC, Bongarzone S, Gee A, et al. PET imaging as a tool for affine registration of brain images. Med Image Anal 2001;5:143–56 assessing COVID-19 brain changes. Trends Neurosci 2020;43:935– CrossRef Medline 38 CrossRef Medline 28. Penny WD, Friston KJ, Ashburner JT, et al. Statistical Parametric 16. Rodríguez-Alfonso B, Solís SR, Silva-Hernández L, et al. F-FDG- Mapping: The Analysis of Functional Brain Images. Elsevier; 2011 PET/CT in SARS-CoV-2 infection and its sequelae. Rev Esp Med 29. Rosa PA, Cerami C, Gallivanone F, et al; EADC-PET Consortium. A Nucl Imagen Mol (Engl Ed) 2021;40:299–309 CrossRef Medline standardized [18F]-FDG-PET template for spatial normalization 17. Martini AL, Carli G, Kiferle L, et al. Time-dependent recovery of in statistical parametric mapping of dementia. Neuroinformatics brain hypometabolism in neuro-COVID-19 patients. Eur J Nucl 2014;12:575–93 CrossRef Medline 30. Meyer PT, Hellwig S, Blazhenets G, et al. Molecular imaging find- Med Mol Imaging 2022;50:90–102 CrossRef Medline ings on acute and long-term effects of COVID-19 on the brain: a 18. Kas A, Soret M, Pyatigoskaya N, et al; on the behalf of CoCo- systematic review. JNucl Med 2022;63:971–80 CrossRef Medline Neurosciences study group and COVID SMIT PSL study group. The 31. Niesen M, Trotta N, Noel A, et al. Structural and metabolic brain cerebral network of COVID-19-related encephalopathy: a longitu- abnormalities in COVID-19 patients with sudden loss of smell. dinal voxel-based 18F-FDG-PET study. Eur J Nucl Med Mol Eur J Nucl Med Mol Imaging 2021;48:1890–901 CrossRef Medline Imaging 2021;48:2543–57 CrossRef Medline 32. Starke KR, Reissig D, Petereit-Haack G, et al. The isolated effect of age 19. Guedj E, Campion JY, Dudouet P, et al. (18)F-FDG brain PET hypo- on the risk of COVID-19 severe outcomes: a systematic review with metabolism in patients with long COVID. Eur J Nucl Med Mol meta-analysis. BMJ Glob Health 2021;6:e006434 CrossRef Medline Imaging 2021;48:2823–33 CrossRef Medline 33. Statsenko Y, Zahmi FA, Habuza T, et al. Impact of age and sex on 20. Sollini M, Morbelli S, Ciccarelli M, et al. Long COVID hallmarks on COVID-19 severity assessed from radiologic and clinical findings. [18F]FDG-PET/CT: a case-control study. Eur J Nucl Med Mol Front Cell Infect Microbiol 2022;11:777070 CrossRef Medline Imaging 2021;48:3187–97 CrossRef Medline 34. Goehringer F, Bruyere A, Doyen M, et al. Brain 18F-FDG PET imag- 21. Gandhi RT, Lynch JB, Rio CD. Mild or moderate Covid-19. NEngl J ing in outpatients with post-COVID-19 conditions: findings and Med 2020;383:1757–66 CrossRef Medline associations with clinical characteristics. Eur J Nucl Med Mol 22. Dietz M, Chironi G, Claessens YE, et al; MONACOVID Group. Imaging 2023;50:1084–89 CrossRef Medline COVID-19 pneumonia: relationship between inflammation assessed 35. Hosp JA, Dressing A, Blazhenets G, et al. Cognitive impairment and by whole-body FDG PET/CT and short-term clinical outcome. Eur J altered cerebral glucose metabolism in the subacute stage of Nucl Med Mol Imaging 2021;48:260–68 CrossRef Medline COVID-19. Brain 2021;144:1263–76 CrossRef Medline 522 Debs May 2023 www.ajnr.org
Journal
American Journal of Neuroradiology – American Journal of Neuroradiology
Published: May 1, 2023