Automated Prediction of the Response to Neoadjuvant Chemoradiotherapy in Patients Affected by Rectal Cancer

Abstract

Background: Rectal cancer is a malignant neoplasm of the large intestine resulting from the uncontrolled proliferation of the rectal tract. Predicting the pathologic response of neoadjuvant chemoradiotherapy at an MRI primary staging scan in patients affected by locally advanced rectal cancer (LARC) could lead to significant improvement in the survival and quality of life of the patients. In this study, the possibility of automatizing this estimation from a primary staging MRI scan, using a fully automated artificial intelligence-based model for the segmentation and consequent characterization of the tumor areas using radiomic features was evaluated. The TRG score was used to evaluate the clinical outcome.

Methods: Forty-three patients under treatment in the IRCCS Sant'Orsola-Malpighi Polyclinic were retrospectively selected for the study; a U-Net model was trained for the automated segmentation of the tumor areas; the radiomic features were collected and used to predict the tumor regression grade (TRG) score.

Results: The segmentation of tumor areas outperformed the state-of-the-art results in terms of the Dice score coefficient or was comparable to them but with the advantage of considering mucinous cases. Analysis of the radiomic features extracted from the lesion areas allowed us to predict the TRG score, with the results agreeing with the state-of-the-art results.

Conclusions: The results obtained regarding TRG prediction using the proposed fully automated pipeline prove its possible usage as a viable decision support system for radiologists in clinical practice.

Keywords: artificial intelligence; machine and deep learning; medical imaging; radiomics.

Figure 1

Schematic representation of the proposed pipeline. From the top left: raw T₂-weighted MRI scan; pre-processed image using denoising algorithm and gamma correction for the remotion of possible confounders; segmentation of the lesion areas using the authors’ CNN model; extraction of the radiomic features from the areas identified by the CNN model; and prediction of the TRG score.

Figure 2

Comparison between the ground truth and the results obtained by the proposed pipeline for the lesion segmentation. (a–d) Original MRI scans for adenocarcinoma and mucinous cases. (b–e) Ground truth obtained by manual segmentation performed by experts. (c–f) Predicted segmentation obtained by the proposed U-Net model.

Figure 3

Ranking of the top informative features identified by the proposed pipeline. For each principal component (on the right), the original feature (on the left) that contributes the most to the related principal component is reported.