18F-FDG PET/CT Uptake as a Predictor of Tumor Response in Neoadjuvant Treatment of Rectal Cancer with Dose Escalation using Simultaneous Integrated Boost

Abstract

To assess the value of 18F-FDG PET SUVmax to predict tumor response in neoadjuvant treatment of locally advanced rectal cancer using dose escalation with simultaneous integrated boost (SIB). Forty-four patients with stage II-III rectal adenocarcinoma treated with neoadjuvant dose escalated radiotherapy using SIB and have pre- and post-radiotherapy 18F-FDG PET/CT images were included retrospectively. Median age was 60 (18-81). Female/male ratio was 15/29. Tumor location was distal rectum in 17 patients, mid-rectum in 9 and proximal rectum in 18. Radiotherapy (RT) dose was 50.4 Gy to the pelvic lymph nodes and simultaneously delivered 56 Gy to the tumor, in 28 fractions, using SIB with VMAT, concomitant with capecitabine, 825 mg/m2 bid. Median time to post-RT PET was 9 (5-20) weeks. Thirteen (29.5%) patients refused surgery or were medically inoperable and 31 (70.5%) patients underwent surgery. The ability of pre- and post-RT SUVmax to predict treatment response was analyzed by a Receiver Operating Characteristics (ROC) curve analysis. While evaluating the area under the curve, a 5% type-1 error level was accepted a statistically significant predictive value of the test variables. The relation between SUVmax and tumor response was analyzed using Fisher's exact test. Clinical (n=12) or pathological complete (n=6) or near complete (n=5) response was observed in 23 (52.3%) patients. Twenty-one (47.7%) patients had partial (n=13) or minimal (n=8) response. Median tumor pre- and post-RT SUVmax were 19 (6-37) and 5 (2-13), respectively. Area under the ROC curve was 0.549, and the best cut-off pre-RT SUVmax to predict tumor response was 8.3. Area under the ROC curve was 0.724, and the best cut-off post-RT SUVmax to predict tumor response was 3.3. Post-RT SUVmax <3.3 was significantly associated with complete or near complete response (p=0.017). Post-RT SUVmax was <3.3 in 12 (complete or near complete response: 10, partial or no response: 2) and was >3.3 in 32 (complete or near complete response: 13, partial or no response: 19) patients. However, pre-RT SUVmax did not have any predictive value (p>0.05). Post-RT SUVmax may predict complete or near complete tumor response in rectal adenocarcinoma patients treated with neoadjuvant radiotherapy using dose escalation with SIB.