Does the computed tomography perfusion imaging improve the diagnostic accuracy in the response evaluation of esophageal carcinoma to the neoadjuvant chemoradiotherapy? Preliminary study

Abstract

Purpose: To estimate whether the computed tomography (CT) perfusion imaging could be useful to predict the pathological complete response (pCR) of esophageal cancer to the neoadjuvant chemoradiotherapy (NACRT).

Methods: Twenty-seven patients with the advanced squamous cell esophageal carcinoma, who were treated with concomitant CRT (CIS/5-FU/LV and 45-50 Gy total radiation dose), were re-evaluated using CT examination, which included the low-dose CT perfusion study. CT perfusion series were analysed using the deconvolution-based CT perfusion software (Perfusion 3.0, GE), and color parametric maps of the blood flow (BF), blood volume (BV), mean transit time (MTT), and permeability surface area product (PS) were displayed. All patients were operated and histopathological analysis of the resected esophagus considered the gold standard for pathologic complete response (pCR).

Results: BFpost-NACRT, BVpost-NACRT, and PSpost-NACRT were significantly lower, and MTTpost-NACRT significantly higher in the pCR group. Mean (±SD), or median perfusion parameter values in the pCRs (11 patients) vs non-pCRs (16 patients) were: BFpost-NACRT- 21.4±5.0 vs 86.0±29 ml/min/100 g (p<0.001), BVpost-NACRT- 1.3 vs 3.9 ml/100 g (p<0.001), MTTpost-NACRT- 5.5 vs 3.7 s (p=0.018), and PSpost-NACRT- 5.9 vs 9.8 ml/min/100 g (p=0.006). ROC analysis revealed that BFpost- NACRT (AUC=1.000), BVpost-NACRT (AUC=0.932), MTTpost-NACRT (AUC=0.801), and PSpost-NACRT (AUC=0.844) could predict the pCR (p<0.01), while maximal esophageal wall thickness could not (AUC=0.676, p=0.126). If we set a cut-off value of BFpost-NACRT<30.0 ml/min/100 g, pCR was predicted with sensitivity and specificity of 100%.

Conclusion: CT perfusion imaging enables accurate prediction of pCR of esophageal carcinoma to neoadjuvant chemoradiotherapy.