Sequential FDG-PET/CT reliably predicts response of locally advanced rectal cancer to neo-adjuvant chemo-radiation therapy

Prediction of rectal cancer response to preoperative, neo-adjuvant chemo-radiation therapy (CRT) provides the opportunity to identify patients in whom a major response is expected and who may therefore benefit from alternative surgical approaches. Traditional morphological imaging techniques are effective in defining tumour extension in the initial diagnostic and staging work-up, but perform poorly in distinguishing residual neoplastic tissue from scarring post CRT, when restaging the patient before surgery. Fluorine-18 fluorodeoxyglucose positron emission tomography (FDG-PET) is a promising tool for monitoring the effect of anti-tumour therapy. The aim of this study was to prospectively assess the value of sequential FDG-PET scans in predicting the response of locally advanced rectal cancer to neo-adjuvant CRT.

Forty-four consecutive patients with locally advanced (cT3–4) primary rectal cancer and four patients with pelvic recurrence of rectal cancer were enrolled in this prospective study. Treatment consisted of external beam intensified radiotherapy (50 Gy to the posterior pelvis, 56 Gy to the tumour), chemotherapy (in most cases PVI 5-FU at 300 mg/m² per day) and, 8–10 weeks later, surgery with curative intent. All patients underwent FDG-PET/CT both before CRT and 5–6 weeks after completing CRT. One patient died before surgery because of acute myocardial infarction, and was therefore excluded from further analysis. Semi-quantitative measurements of FDG uptake (SUV_max), absolute difference (ΔSUV_max) and percent SUV_max difference (Response Index, RI) between pre- and post-CRT PET scans were considered. Results were correlated with pathological response, assessed both by histopathological staging of the surgical specimens (pTNM) and by the tumour regression grade (TRG) according to Mandard’s criteria (patients with TRG1–2 being defined as responders and patients with TRG3–5 as non-responders).

Following neo-adjuvant CRT, of the 45 patients submitted to surgery, 23 (51.1%) were classified as responders according to Mandard’s criteria (8 TRG1 and 15 TRG2), while the remaining 22 (48.9%) were non-responders (9 TRG3 and 13 TRG4–5). Considering all patients, the mean pre-CRT SUV_max was 15.6, significantly higher than the mean value of 5.4 post CRT (p < 0.001). Nevertheless, when stratifying patients according to response to CRT (using Mandard’s criteria), the mean RI was significantly higher in responders than in non-responders (75.9% versus 46.9%,p = 0.0015). Using a 66.2% SUV_max decrease as the cut-off value (identified by ROC analysis) for defining response to therapy, the following parameters were obtained: 79.2% specificity, 81.2% sensitivity, 77% positive predictive value, 89% negative predictive value and 80% overall accuracy.

The results suggest the potential utility of FDG-PET as a complementary diagnostic and prognostic procedure in the assessment of neo-adjuvant CRT response of locally advanced rectal cancer. ΔSUV_max and RI seem the best predictors of CRT response.