Purpose or Learning Objective To investigate whether in locally advanced rectal cancer (LARC) treated with neo-adjuvant chemoradiotherapy (nCRT), radiomics on T2 weighted (T2w) MRI sequences can discriminate responder (R) and non-responder (NR) patients based on the Tumour Regression Grade (TRG) assigned after surgical resection Methods or Background This study retrospectively enrols 40 patients undergoing pre-therapy 1.5T-MRI. Regions of Interest (ROIs) are manually outlined in all slices of the tumour’s site on T2w sequences in the oblique-axial plane, acquired with 3 mm slice thickness. Based on TRG, R patients have complete and partial nCRT response (TRG=[0,1], n°15) while NR patients have a minimal and poor nCRT response (TRG=[2,3], n°25). Eighty-four local first-order radiomic features (RFs) are extracted from tumour ROIs. To prevent overfitting, only single RFs are investigated to discriminate Rs and NRs. The most performing feature is selected through a univariate analysis guided by one-tail Wilcoxon rank-sum test (p=0.05 significance level). To assess the feature discrimination capability, ROC curve analysis is performed, through AUC computation, Youden Index (YI) for sensitivity and specificity. Results or Findings One RF measuring the local heterogeneity of T2w values within tumour ROIs discriminates Rs and NRs with p~10-5, AUC=0.90 (95%CI, 0.73-0.96), with YI=0.68 corresponding to sensitivity=80% and specificity=88%. The separation achieved highlights 3 false positives and 3 false negatives. Conclusion Pre-therapy baseline tumour heterogeneity measured from T2w-MR images has a very promising role in predicting the TRG histological classification. Patients with lower tumour heterogeneity at pre-therapy show a better response to nCRT. Limitations This study involves a small number of patients. However, one-only feature is considered and such a strong discrimination stresses the future role of the feature in a classification study.

Locally advanced rectal cancer: T2w-MRI-based radiomics may detect responder patients undergoing neoadjuvant chemoradiotherapy

Alessandro. Bevilacqua;Margherita Mottola;Silvia Lo Monaco;Arrigo Cattabriga;Maria Adriana Cocozza;Dajana Cuicchi;Luigi Ricciardiello;Rita Golfieri
2021

Abstract

Purpose or Learning Objective To investigate whether in locally advanced rectal cancer (LARC) treated with neo-adjuvant chemoradiotherapy (nCRT), radiomics on T2 weighted (T2w) MRI sequences can discriminate responder (R) and non-responder (NR) patients based on the Tumour Regression Grade (TRG) assigned after surgical resection Methods or Background This study retrospectively enrols 40 patients undergoing pre-therapy 1.5T-MRI. Regions of Interest (ROIs) are manually outlined in all slices of the tumour’s site on T2w sequences in the oblique-axial plane, acquired with 3 mm slice thickness. Based on TRG, R patients have complete and partial nCRT response (TRG=[0,1], n°15) while NR patients have a minimal and poor nCRT response (TRG=[2,3], n°25). Eighty-four local first-order radiomic features (RFs) are extracted from tumour ROIs. To prevent overfitting, only single RFs are investigated to discriminate Rs and NRs. The most performing feature is selected through a univariate analysis guided by one-tail Wilcoxon rank-sum test (p=0.05 significance level). To assess the feature discrimination capability, ROC curve analysis is performed, through AUC computation, Youden Index (YI) for sensitivity and specificity. Results or Findings One RF measuring the local heterogeneity of T2w values within tumour ROIs discriminates Rs and NRs with p~10-5, AUC=0.90 (95%CI, 0.73-0.96), with YI=0.68 corresponding to sensitivity=80% and specificity=88%. The separation achieved highlights 3 false positives and 3 false negatives. Conclusion Pre-therapy baseline tumour heterogeneity measured from T2w-MR images has a very promising role in predicting the TRG histological classification. Patients with lower tumour heterogeneity at pre-therapy show a better response to nCRT. Limitations This study involves a small number of patients. However, one-only feature is considered and such a strong discrimination stresses the future role of the feature in a classification study.
ECR 2021 – BOOK OF ABSTRACTS
1
1
Alessandro. Bevilacqua, Francesca Coppola, Margherita Mottola, Silvia Lo Monaco, Arrigo Cattabriga, Maria Adriana Cocozza, Dajana Cuicchi, Luigi Ricciardiello, Rita Golfieri
File in questo prodotto:
Eventuali allegati, non sono esposti

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: http://hdl.handle.net/11585/785455
 Attenzione

Attenzione! I dati visualizzati non sono stati sottoposti a validazione da parte dell'ateneo

Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus ND
  • ???jsp.display-item.citation.isi??? ND
social impact