Carlo N De Cecco1,2, Maria Ciolina1, Damiano Caruso1, Marco Rengo1, Balaji Ganeshan3, Felix G Meinel4, Daniela Musio5, Francesca De Felice5, Vincenzo Tombolini5, Andrea Laghi6. 1. Diagnostic Imaging Unit, Department of Radiological Sciences, Oncology and Pathology, I.C.O.T. Hospital, University of Rome "Sapienza" - Polo Pontino, Latina, Italy. 2. Department of Radiology, Medical University of South Carolina, Charleston, SC, USA. 3. Clinical Imaging Sciences Centre, Brighton and Sussex Medical School, Falmer, Sussex, United Kingdom. 4. Institute for Clinical Radiology, Ludwig-Maximilians-University Hospital, Munich, Germany. 5. Department of Radiotherapy, University of Rome "Sapienza", Rome, Italy. 6. Diagnostic Imaging Unit, Department of Radiological Sciences, Oncology and Pathology, I.C.O.T. Hospital, University of Rome "Sapienza" - Polo Pontino, Latina, Italy. andrea.laghi@uniroma1.it.
Abstract
PURPOSE: To determine the performance of texture analysis (TA), diffusion-weighted imaging, and perfusion MR (pMRI) in predicting tumoral response in patients treated with neoadjuvant chemoradiotherapy (CRT). METHODS: 12 consecutive patients (8 females, 4 males, 63.2 ± 13.4 years) with rectal cancer were prospectively enrolled, and underwent pre-treatment 3T MRI. Treatment protocol consisted of neoadjuvant CRT with oxaliplatin and 5-fluorouracile. Unenhanced T2-weighted images TA (kurtosis), apparent diffusion coefficient (ADC), and pMRI parameters (Ktrans, Kep, Ve, IAUGC) were quantified by manually delineating a region of interest around the tumor outline. After CRT, all patients underwent complete surgical resection and the surgical specimen served as the gold standard. Receiver operating characteristic (ROC) curve analysis was performed to assess the discriminatory power of each quantitative parameter to predict complete response. RESULTS: Pathological complete response (pCR) was reported in six patients and partial response (PR) in three patients. Three patients were classified as non-responders (NR). Pre-treatment kurtosis was significantly lower in the pCR sub-group in comparison with PR + NR (p = .01). Among ADC and pMRI parameters, only Ve was significantly lower in the pCR sub-group compared with PR + NR (p = .01). A significant negative correlation between kurtosis and ADC (r = -0.650, p = .022) was observed. Pre-treatment area under the ROC curves (AUC), to discriminate between pCR and PR + NR, was significantly higher for kurtosis (0.861, p = .001) and Ve (0.861, p = .003) compared to all other parameters. The optimal cutoff value for pre-treatment kurtosis and Ve was ≤0.19 (100% sensitivity, 67% specificity) and ≤0.311 (83% sensitivity, 83% specificity), respectively. CONCLUSION: Pre-treatment kurtosis derived from T2w images and Ve from pMRI have the potential to act as imaging biomarkers of rectal cancer response to neoadjuvant CRT.
PURPOSE: To determine the performance of texture analysis (TA), diffusion-weighted imaging, and perfusion MR (pMRI) in predicting tumoral response in patients treated with neoadjuvant chemoradiotherapy (CRT). METHODS: 12 consecutive patients (8 females, 4 males, 63.2 ± 13.4 years) with rectal cancer were prospectively enrolled, and underwent pre-treatment 3T MRI. Treatment protocol consisted of neoadjuvant CRT with oxaliplatin and 5-fluorouracile. Unenhanced T2-weighted images TA (kurtosis), apparent diffusion coefficient (ADC), and pMRI parameters (Ktrans, Kep, Ve, IAUGC) were quantified by manually delineating a region of interest around the tumor outline. After CRT, all patients underwent complete surgical resection and the surgical specimen served as the gold standard. Receiver operating characteristic (ROC) curve analysis was performed to assess the discriminatory power of each quantitative parameter to predict complete response. RESULTS: Pathological complete response (pCR) was reported in six patients and partial response (PR) in three patients. Three patients were classified as non-responders (NR). Pre-treatment kurtosis was significantly lower in the pCR sub-group in comparison with PR + NR (p = .01). Among ADC and pMRI parameters, only Ve was significantly lower in the pCR sub-group compared with PR + NR (p = .01). A significant negative correlation between kurtosis and ADC (r = -0.650, p = .022) was observed. Pre-treatment area under the ROC curves (AUC), to discriminate between pCR and PR + NR, was significantly higher for kurtosis (0.861, p = .001) and Ve (0.861, p = .003) compared to all other parameters. The optimal cutoff value for pre-treatment kurtosis and Ve was ≤0.19 (100% sensitivity, 67% specificity) and ≤0.311 (83% sensitivity, 83% specificity), respectively. CONCLUSION: Pre-treatment kurtosis derived from T2w images and Ve from pMRI have the potential to act as imaging biomarkers of rectal cancer response to neoadjuvant CRT.
Authors: Jacob Antunes; Satish Viswanath; Justin T Brady; Benjamin Crawshaw; Pablo Ros; Scott Steele; Conor P Delaney; Raj Paspulati; Joseph Willis; Anant Madabhushi Journal: Acad Radiol Date: 2018-01-19 Impact factor: 3.173