Image intensity histograms as imaging biomarkers: application to immune-related colitis

Daniel T Huff, Peter Ferjancic, Mauro Namías, Hamid Emamekhoo, Scott B Perlman, Robert Jeraj, Daniel T Huff, Peter Ferjancic, Mauro Namías, Hamid Emamekhoo, Scott B Perlman, Robert Jeraj

Abstract

Purpose.To investigate image intensity histograms as a potential source of useful imaging biomarkers in both a clinical example of detecting immune-related colitis (irColitis) in18F-FDG PET/CT images of immunotherapy patients and an idealized case of classifying digital reference objects (DRO).Methods.Retrospective analysis of bowel18F-FDG uptake in N = 40 patients receiving immune checkpoint inhibitors was conducted. A CNN trained to segment the bowel was used to generate the histogram of bowel18F-FDG uptake, and percentiles of the histogram were considered as potential metrics for detecting inflammation associated with irColitis. A model of the colon was also considered using cylindrical DRO. Classification of DRO with different intensity distributions was undertaken under varying geometry and noise settings.Results.The most predictive biomarker of irColitis was the 95th percentile of the bowel SUV histogram (SUV95%). Patients later diagnosed with irColitis had a significantly higher increase in SUV95%from baseline to first on-treatment PET than patients who did not experience irColitis (p = 0.02). An increase in SUV95%> + 40% separated pre-irColitis change from normal variability with a sensitivity of 75% and specificity of 88%. Furthermore, histogram percentiles were ideal metrics for classifying 'hot center' and 'cold center' DRO, and were robust to varying DRO geometry and noise, and to the presence of spoiler volumes unrelated to the detection task.Conclusions.The 95th percentile of the bowel SUV histogram was the optimal metric for detecting irColitis on18F-FDG PET/CT. Image intensity histograms are a promising source of imaging biomarkers for clinical tasks.

Keywords: 18F-FDG PET/CT; adverse event; immunotherapy; segmentation.

© 2021 IOP Publishing Ltd.

Figures

Figure 1.
Figure 1.
Visual description of calculating percentile metrics. (a) The raw image intensity histogram H(x) is constructed by masking the image with the ROI mask. (b) The cumulative density function (cdf) of this histogram is computed and normalized in accordance with equation (2). (c) Percentile metrics PY% can be inferred easily from the final cumulative histogram. As an example, we show how P80% can be read from this plot.
Figure 2.
Figure 2.
(a) Geometry of cylindrical digital reference objects (DRO). A blue ‘background’ cylinder has radius R, height Z, and intensity I0. Adding a red cylinder with radius r, height z, and intensity I is done to create either ‘hot center’ (if I > I0) or ‘cold center’ (if I < I0) DRO. Adding a green spoiler cylinder with radius rs, height zs, and intensity Is to simulate the impact of an image feature unrelated to the classification task. (b) DRO axial slices. Radius of added ‘hot center’ or ‘cold center’ increases bottom to top, contrast between added cylinder and background increases from left to right.
Figure 3.
Figure 3.
DRO classification performance for varying cylinder radii r and height z. For ‘background’ versus ‘hot center’ classification (a), performance saturation occurs at percentiles above the volume fraction of the ‘hot center’ cylinder, as defined by Zsat,HC%. For ‘background’ versus ‘cold cylinder’ classification (b), performance saturation occurs at percentiles below the volume fraction of the ‘cold center’ cylinder, as defined by Zsat,CC%. For ‘background versus ‘hot center’ with spoilers classification (c), performance is highest around Zmin,SP% and drops significantly above Zmax,SP%.
Figure 4.
Figure 4.
DRO classification performance for varying intensity I and noise parameter k for the three DRO classification tasks: (a) background versus hot center, (b) background versus cold center, and (c) background versus hot center with spoiler.
Figure 5.
Figure 5.
Patient change in bowel SUV95%, SUVmean, SUVmax, and SUVtotal from PET1 to PET2 in patients with normal bowel (NB) and patients who later received a diagnosis of irColitis (AE). (a) Change in bowel SUV95% defined as the relative change in the 95th percentile value of the bowel SUV histogram from PET1 to PET2 was found to be significantly higher in patients who later received a diagnosis of irColitis than in patients with normal bowel (Wilcoxon rank-sum test, p = 0.023).
Figure 6.
Figure 6.
(a) Receiver operator characteristic (ROC) curve for change in SUV95% achieved an AUROC of 0.86 for predicting irColitis. Comparison metrics SUVmean, SUVmax, and SUVtotal were not predictive of irColitis. An increase in SUV95% of greater than +40% (operating point indicated by *) had a sensitivity of 75% and specificity of 88% in identifying pre-colitis bowel change. (b) Sensitivity of irColitis detection to bowel SUV histogram percentile. The AUROC was calculated for each percentile of the bowel SUV histogram X (SUVX%) where X was varied from 0 to 100. Optimal performance was observed at SUV95%, with an AUC of 0.86. Percentiles from SUV93% to SUV98% all achieved AUC > 0.80.
Figure 7.
Figure 7.
Example bowel SUV histograms. For a patient with no clinical diagnosis of irColitis (left), the histogram at baseline (PET1) and at first follow-up (PET2) are nearly identical and change in SUV95% is small. However, for a patient later diagnosed with irColitis (right), histogram values at high percentiles (>80th percentile) are markedly higher at PET2 than at PET1. The relative increase in SUV95% for this patient is +55%.
Figure 8.
Figure 8.
Longitudinal time series relative to treatment start (day 0) of bowel SUV95% for four patients who experienced irColitis (colored lines). The shaded gray region indicates the 95% confidence interval for normalized change in bowel SUV95% for patients with normal bowel (NB) who did not experience irColitis. The confidence interval was constructed as [x¯−1.96σn,x¯+1.96σn] where x¯ is the sample mean, σ is the sample standard deviation, and n is the number of observations (n = 33). The dotted line indicates the optimal threshold identified through ROC analysis (ΔSUV95% > +40%). Arrows indicate when clinical diagnosis of irColitis was made for each patient.
Figure 9.
Figure 9.
Serial 18F-FDG PET maximum intensity projections of a 68-year-old female with metastatic melanoma receiving combination ipilimumab and nivolumab immunotherapy. (a) At baseline 11 days before treatment start, disease in the pelvis and lower extremities is seen and bowel uptake is normal. (b) Day 84 after treatment start, near complete response of disease sites in the pelvis and lower leg, and moderate increase in bowel uptake can be seen (ΔSUV95% from PET1 to PET2 was +59%). (c) Day 173, continued response, marked elevated bowel uptake is apparent. On day 195 (between c and d), the patient was hospitalized with blood in their stool (Grade 3 Colitis). Colonoscopy with biopsy confirmed irColitis. The patient also had Grade 3 rash during this time and was started on systemic steroids 3 weeks before time point c. (d) Day 273, continued response, elevated bowel uptake remains. Diffuse, elevated lung uptake is apparent as well. Patient received a diagnosis of immune-related pneumonitis. (e) Day 399, continued response, partial resolution of elevated bowel uptake and complete resolution of elevated lung uptake is seen. The patient’s irColitis resolved after completion of slow tapered course of steroids in addition to a course of Budesonide. Scans (a)–(d) were all taken on the same PET/CT scanner (D710). Scan (e) was taken on a different scanner (MI).
Figure 10.
Figure 10.
Effect of harmonization on bowel SUV metrics extracted from PET images. Baseline (PET1) SUVmean, SUVtotal, SUV95%, and change in SUV95% from PET1 to PET2 (ΔSUV95%) are largely unaffected by harmonization, while SUVmax tends to be decreased by harmonization, especially for high SUVmax values. This behavior is due to the optimal filtering approach to harmonization we employed (Namías et al 2020), which smooths down high SUV values.

References

    1. Aerts HJ. et al. Decoding tumour phenotype by noninvasive imaging using a quantitative radiomics approach. Nat. Commun. 2014;5:4006.
    1. Baiocco S, Matteucci F, Mezzenga E, Caroli P, Di Iorio V, Cittanti C, Bevilacqua A, Paganelli G and Sarnelli A 2019. SUV 95th as a reliable alternative to SUV max for determining renal uptake in [68 Ga] PSMA PET/CT Molecular Imaging and Biology 1–8
    1. Beck KE, Blansfield JA, Tran KQ, Feldman AL, Hughes MS, Royal RE, Kammula US, Topalian SL, Sherry RM and Kleiner D 2006. Enterocolitis in patients with cancer after antibody blockade of cytotoxic T-lymphocyte–associated antigen 4 J Clin Oncol 24 2283.
    1. Chen R, Chen Y, Huang G and Liu J 2019. Relationship between PD-L1 expression and (18)F-FDG uptake in gastric cancer Aging (Albany NY) 11 12270–7
    1. Cho SY, Huff DT, Jeraj R and Albertini MR 2020. FDG PET/CT for assessment of immune therapy: opportunities and understanding pitfalls Semin. Nucl. Med 50 518–31
    1. Crombé A, Périer C, Kind M, De Senneville BD, Le Loarer F, Italiano A, Buy X and Saut O 2019. T2-based MRI Delta- radiomics improve response prediction in soft-tissue sarcomas treated by neoadjuvant chemotherapy J. Magn. Reson. Imaging 50 497–510
    1. Drzymala R, Mohan R, Brewster L, Chu J, Goitein M, Harms W and Urie M 1991. Dose-volume histograms International Journal of Radiation Oncology* Biology* Physics 21 71–8
    1. Fedorov A, Beichel R, Kalpathy-Cramer J, Finet J, Fillion-Robin J-C, Pujol S, Bauer C, Jennings D, Fennessy F and Sonka M 2012. 3D slicer as an image computing platform for the quantitative imaging network J. Magn. Reson. Imaging 30 1323–41
    1. Gillies RJ, Kinahan PE and Hricak H 2016. Radiomics: images are more than pictures, they are data Radiology 278 563–77
    1. Graffy PM, Liu J, O’Connor S, Summers RM and Pickhardt PJ 2019. Automated segmentation and quantification of aortic calcification at abdominal CT: application of a deep learning-based algorithm to a longitudinal screening cohort Abdominal Radiology 44 2921–8
    1. Heimann T, Van Ginneken B, Styner MA, Arzhaeva Y, Aurich V, Bauer C, Beck A, Becker C, Beichel R and Bekes G 2009. Comparison and evaluation of methods for liver segmentation from CT datasets IEEE Trans. Med. Imaging 28 1251–65
    1. Horne ZD, Clump DA, Vargo JA, Shah S, Beriwal S, Burton SA, Quinn AE, Schuchert MJ, Landreneau RJ and Christie NA 2014. Pretreatment SUV max predicts progression-free survival in early-stage non-small cell lung cancer treated with stereotactic body radiation therapy Radiation Oncology 9 1–6
    1. Iravani A and Hicks RJ 2020. Atlas of Response to Immunotherapy ed Lopci E and Fanti S (Cham: Springer International Publishing; ) pp 101–15
    1. Jimenez-del-Toro O, Müller H, Krenn M, Gruenberg K, Taha AA, Winterstein M, Eggel I, Foncubierta-Rodríguez A, Goksel O and Jakab A 2016. Cloud-based evaluation of anatomical structure segmentation and landmark detection algorithms: VISCERAL anatomy benchmarks IEEE Trans. Med. Imaging 35 2459–75
    1. Kamnitsas K, Ledig C, Newcombe VF, Simpson JP, Kane AD, Menon DK, Rueckert D and Glocker B 2017. Efficient multiscale 3D CNN with fully connected CRF for accurate brain lesion segmentation Med. Image Anal. 36 61–78
    1. Kim et al. 2014. Quantitative analysis of computed tomography images and early detection of cerebral edema for pediatric traumatic brain injury patients: retrospective study BMC medicine 12 1–16
    1. Kinahan P, Muzi M, Bialecki B, Herman B and Coombs L 2019. ACRIN 6668 trial NSCLC-FDG-PET Data from ()
    1. Lang N, Dick J, Slynko A, Schulz C, Dimitrakopoulou-Strauss A, Sachpekidis C, Enk AH and Hassel JC 2019. Clinical significance of signs of autoimmune colitis in 18F-fluorodeoxyglucose positron emission tomography-computed tomography of 100 stage-IV melanoma patients Immunotherapy 11 667–76
    1. Larkin J, Chiarion-Sileni V, Gonzalez R, Grob JJ, Cowey CL, Lao CD, Schadendorf D, Dummer R, Smylie M and Rutkowski P 2015. Combined nivolumab and ipilimumab or monotherapy in untreated melanoma New Engl. J. Med 373 23–34
    1. Liu Y, Lei Y, Fu Y, Wang T, Tang X, Jiang X, Curran WJ, Liu T, Patel P and Yang X 2020. CT-based multi-organ segmentation using a 3D self-attention U-net network for pancreatic radiotherapy Med. Phys 47 4316–24
    1. Mu W, Tunali I, Gray JE, Qi J, Schabath MB and Gillies RJ 2019. Radiomics of (18)F-FDG PET/CT images predicts clinical benefit of advanced NSCLC patients to checkpoint blockade immunotherapy Eur. J. Nucl. Med. Mol. Imaging 47 1168–92
    1. Nakamura K, Kodama J, Okumura Y, Hongo A, Kanazawa S and Hiramatsu Y 2010. The SUVmax of 18F-FDG PET correlates with histological grade in endometrial cancer International Journal of Gynecologic Cancer 20 100–15
    1. Namías M, Huff D, Weisman A, Bradshaw T and Jeraj R 2020. Retrospective quantitative harmonization in PET using deconvolution and optimal filtering Bull. Am. Phys. Soc 65
    1. Nie K, Shi L, Chen Q, Hu X, Jabbour SK, Yue N, Niu T and Sun X 2016. Rectal cancer: assessment of neoadjuvant chemoradiation outcome based on radiomics of multiparametric MRI Clinical cancer research 22 5256–64
    1. O’Connor JP, Rose CJ, Waterton JC, Carano RA, Parker GJ and Jackson A 2015. Imaging intratumor heterogeneity: role in therapy response, resistance, and clinical outcome Clinical Cancer Research 21 249–57
    1. Pauly O, Ahmadi S-A, Plate A, Boetzel K and Navab N 2012. Int. conf. on medical image computing and computer-assisted intervention vol. Series (Berlin: Springer; ) pp 443–50
    1. Pillai SC, Gill D, Webster R, Howman-Giles R and Dale RC 2010. Cortical hypometabolism demonstrated by PET in relapsing NMDA receptor encephalitis Pediatric neurology 43 217–20
    1. Puzanov I, Diab A, Abdallah K, Bingham C, Brogdon C, Dadu R, Hamad L, Kim S, Lacouture M and LeBoeuf N 2017. Managing toxicities associated with immune checkpoint inhibitors: consensus recommendations from the society for immunotherapy of cancer (SITC) toxicity management working group J Immunother Cancer 5 95.
    1. Rodrigues G, Lock M, D’Souza D, Yu E and Van Dyk J 2004. Prediction of radiation pneumonitis by dose–volume histogram parameters in lung cancer—a systematic review Radiother. Oncol 71 127–38
    1. Sosa A, Lopez Cadena E, Simon Olive C, Karachaliou N and Rosell R 2018. Clinical assessment of immune-related adverse events Ther Adv Med Oncol 10 1758835918764628
    1. Sun R et al. 2018. A radiomics approach to assess tumour-infiltrating CD8 cells and response to anti-PD-1 or anti-PD-L1 immunotherapy: an imaging biomarker, retrospective multicohort study Lancet Oncol. 19 1180–91
    1. Tieleman T and Hinton G 2012. Lecture 6.5-rmsprop: divide the gradient by a running average of its recent magnitude COURSERA: Neural networks for machine learning 4 26–31
    1. Tofts PS, Benton CE, Weil RS, Tozer DJ, Altmann DR, Jäger HR, Waldman AD and Rees JH 2007. Quantitative analysis of whole-tumor Gd enhancement histograms predicts malignant transformation in low-grade gliomas Journal of Magnetic Resonance Imaging: An Official Journal of the International Society for Magnetic Resonance in Medicine 25 208–14
    1. van Velden FH, Cheebsumon P, Yaqub M, Smit EF, Hoekstra OS, Lammertsma AA and Boellaard R 2011. Evaluation of a cumulative SUV-volume histogram method for parameterizing heterogeneous intratumoural FDG uptake in non-small cell lung cancer PET studies Eur. J. Nucl. Med. Mol. Imaging 38 1636–47
    1. Wachsmann JW, Ganti R and Peng F 2017. Immune-mediated disease in ipilimumab immunotherapy of melanoma with FDG PET-CT Acad. Radiol 24 111–5
    1. Wang G, He L, Yuan C, Huang Y, Liu Z and Liang C 2018. Pretreatment MR imaging radiomics signatures for response prediction to induction chemotherapy in patients with nasopharyngeal carcinoma Eur. J. Radiol 98 100–6

Source: PubMed

Sponsors and Collaborators

Medical Conditions

Drug Interventions

3
Subscribe