Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2022 Jun 16:12:868265.
doi: 10.3389/fonc.2022.868265. eCollection 2022.

Clinically Interpretable Radiomics-Based Prediction of Histopathologic Response to Neoadjuvant Chemotherapy in High-Grade Serous Ovarian Carcinoma

Leonardo Rundo  1   2 , Lucian Beer  1   2   3 , Lorena Escudero Sanchez  1   2 , Mireia Crispin-Ortuzar  2   4   5 , Marika Reinius  2   4   6 , Cathal McCague  1   2 , Hilal Sahin  1   2   7 , Vlad Bura  1   2   8 , Roxana Pintican  8   9 , Marta Zerunian  10 , Stephan Ursprung  1 , Iris Allajbeu  1   6 , Helen Addley  6 , Paula Martin-Gonzalez  2   4 , Thomas Buddenkotte  11 , Naveena Singh  12 , Anju Sahdev  13 , Ionut-Gabriel Funingana  2   4   6 , Mercedes Jimenez-Linan  2   6 , Florian Markowetz  2   4 , James D Brenton  2   4   6 , Evis Sala  1   2   6 , Ramona Woitek  1   2   3
Affiliations

Clinically Interpretable Radiomics-Based Prediction of Histopathologic Response to Neoadjuvant Chemotherapy in High-Grade Serous Ovarian Carcinoma

Leonardo Rundo et al. Front Oncol. .

Abstract

Background: Pathological response to neoadjuvant treatment for patients with high-grade serous ovarian carcinoma (HGSOC) is assessed using the chemotherapy response score (CRS) for omental tumor deposits. The main limitation of CRS is that it requires surgical sampling after initial neoadjuvant chemotherapy (NACT) treatment. Earlier and non-invasive response predictors could improve patient stratification. We developed computed tomography (CT) radiomic measures to predict neoadjuvant response before NACT using CRS as a gold standard.

Methods: Omental CT-based radiomics models, yielding a simplified fully interpretable radiomic signature, were developed using Elastic Net logistic regression and compared to predictions based on omental tumor volume alone. Models were developed on a single institution cohort of neoadjuvant-treated HGSOC (n = 61; 41% complete response to NCT) and tested on an external test cohort (n = 48; 21% complete response).

Results: The performance of the comprehensive radiomics models and the fully interpretable radiomics model was significantly higher than volume-based predictions of response in both the discovery and external test sets when assessed using G-mean (geometric mean of sensitivity and specificity) and NPV, indicating high generalizability and reliability in identifying non-responders when using radiomics. The performance of a fully interpretable model was similar to that of comprehensive radiomics models.

Conclusions: CT-based radiomics allows for predicting response to NACT in a timely manner and without the need for abdominal surgery. Adding pre-NACT radiomics to volumetry improved model performance for predictions of response to NACT in HGSOC and was robust to external testing. A radiomic signature based on five robust predictive features provides improved clinical interpretability and may thus facilitate clinical acceptance and application.

Keywords: chemotherapy response score; computed tomography; neoadjuvant chemotherapy; ovarian cancer; radiomics.

PubMed Disclaimer

Conflict of interest statement

JB is a shareholder of Tailor Bio Ltd, Rutland, United Kingdom; receives honoraria from GlaxoSmithKline, London, United Kingdom and AstraZeneca, Cambridge, United Kingdom; receives research funding from Aprea Therapeutics AB, Massachusetts, United States; and holds patents for methods for predicting treatment response in cancers. ES receives honoraria from GlaxoSmithKline, London, United Kingdom and GE Healthcare, Illinois, United States, and is co-founder and shareholder of Lucida Medical Ltd, Cambridge, United Kingdom. The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Figures

Figure 1
Figure 1
Overall design of the study for identifying radiomic predictors of CRS-confirmed response. Pre- and post-NACT CT images were analyzed. CRS classification is tabulated.
Figure 2
Figure 2
(A) Scheme of the nested k-fold cross-validation (for k outer = 5 and k inner = 5). The nested fitting procedure was repeated 100 times with different random permutations of the discovery dataset. (B) Majority voting for the ensemble of classifiers used for testing on the external test cohort (the dashed red lines denote the decision thresholds optimized according to the inner CV loop). (C) Workflow of the radiomics pipeline for CRS prediction.
Figure 3
Figure 3
Boxplots of the whole tumor and solid tumor volume in patients with non-complete (CRS1-2) and complete response (CRS3) from the (A) discovery (n = 61, non-complete response = 36, complete response = 25) and (B) external test cohorts (n = 48, non-complete response = 38, complete response = 10). Percentage change of whole tumor and solid tumor volume is shown in (C) for the discovery cohort and in (D) for the external test cohort. For pre- and post-NACT volumes, a logarithmic scale was used on the y-axis.
Figure 4
Figure 4
CRS classification results in terms of AUC and G-mean (first row), along with sensitivity and specificity (second row) and PPV and NPV (third row): (A, C, E) discovery cohort; (B, D, F) external test set. We considered the pre-NACT volumetric model and radiomic models fitted on either all the preprocessed features (robust and non-redundant) or only on the most frequently selected (i.e., relevant) features along with omental tumor volume. The variability across 100 repetitions was considered. The dots and error bars denote the average value and the standard deviation, respectively. Brackets denote statistical significance of particular interest using a Wilcoxon rank-sum test. Notation: *p < 0.05, **p < 0.01, ***p < 0.001, ****p ≪ 0.0001.
See this image and copyright information in PMC

References

    1. Vergote I, Tropé CG, Amant F, Kristensen GB, Ehlen T, Johnson N, et al. . Neoadjuvant Chemotherapy or Primary Surgery in Stage IIIC or IV Ovarian Cancer. N Engl J Med (2010) 363:943–53. doi: 10.1056/NEJMoa0908806 - DOI - PubMed
    1. Kehoe S, Hook J, Nankivell M, Jayson GC, Kitchener H, Lopes T, et al. . Primary Chemotherapy Versus Primary Surgery for Newly Diagnosed Advanced Ovarian Cancer (CHORUS): An Open-Label, Randomised, Controlled, non-Inferiority Trial. Lancet (2015) 386:249–57. doi: 10.1016/S0140-6736(14)62223-6 - DOI - PubMed
    1. Coleridge SL, Bryant A, Kehoe S, Morrison J. Neoadjuvant Chemotherapy Before Surgery Versus Surgery Followed by Chemotherapy for Initial Treatment in Advanced Ovarian Epithelial Cancer. Cochrane Database Syst Rev (2021) 7:CD005343. doi: 10.1002/14651858.CD005343.pub6 - DOI - PMC - PubMed
    1. Knisely AT, St Clair CM, Hou JY, Collado FK, Hershman DL, Wright JD, et al. . Trends in Primary Treatment and Median Survival Among Women With Advanced-Stage Epithelial Ovarian Cancer in the US From 2004 to 2016. JAMA Netw Open (2020) 3:e2017517. doi: 10.1001/jamanetworkopen.2020.17517 - DOI - PMC - PubMed
    1. Morgan RD, McNeish IA, Cook AD, James EC, Lord R, Dark G, et al. . Objective Responses to First-Line Neoadjuvant Carboplatin-Paclitaxel Regimens for Ovarian, Fallopian Tube, or Primary Peritoneal Carcinoma (ICON8): Post-Hoc Exploratory Analysis of a Randomised, Phase 3 Trial. Lancet Oncol (2021) 22:277–88. doi: 10.1016/S1470-2045(20)30591-X - DOI - PMC - PubMed

LinkOut - more resources