Predicting pathological response after radio-chemotherapy for rectal cancer: Impact of late oxaliplatin administration

Background and purpose: A previously introduced index based on early tumor (GTV) regression (ERI_TCP) during neo-adjuvant radio-chemotherapy of rectal cancer was used to investigate the impact of changes of oxaliplatin (OXA) delivery on the prediction of pathological complete response (pCR) and residual vital cell (RVC) fraction.

Materials and methods: Ninety-five patients were treated following an adaptive protocol (41.4 Gy/18fr; 2.3 Gy/fr) delivering a simultaneous integrated boost to the residual GTV in the last 6 fractions (3 Gy/fr). OXA was delivered on days -14, 0 (start of RT) and +14. Based on the oncologist's preference, the last OXA cycle was not administered for 36 patients. MRIs taken at planning and at mid-RT were used to calculate ERI_TCP, before the timing of the third OXA cycle. The impact of OXA cycles and the discriminative power of ERI_TCP in predicting the pathological response (pCR, RVC >10%) were quantified. Multivariate logistic regression was performed to assess predictive models.

Results: Two patients with complete clinical remission refused surgery (cCR_ww). Complete post-surgical data of 54/59 and 35/36 patients were available for the two groups (3 vs 2 OXA cycles). pCR/pCR + cCR_ww/RVC >10% rates were 31.5/33.9/27.8% and 14.3/14.3/54.3% respectively (p = 0.01-0.07). ERI_TCP showed high negative predictive value (85-91%) for all end-points. The logistic predictive model for pCR included ERI_TCP (OR: 0.93) and OXA cycles (OR: 3.5), with AUC = 0.78. Internal validation through bootstrap confirmed the robustness of the results.

Conclusions: Late omission of OXA dramatically reduced the pathological response. OXA delivery after the assessment of ERI_TCP significantly influenced the relationship between ERI_TCP and pCR.