Comparison of RECIST, EORTC criteria and PERCIST using MRI and PET/CT for evaluation of early response of neoadjuvant chemotherapy in patients with cervical carcinoma

Objectives: The aim of this study was to compare the RECIST1.1 (Response Evaluation Criteria in Solid Tumors) , EORTC( European Organization for Research and Treatment of Cancer ) criteria and the PERCIST 1.0 (Positron Emission Tomography Response Criteria in Solid Tumors) for response evaluation in patients with cervical carcinoma treated with neoadjuvant chemotherapy.

Methods: A total of 21 patients with cervical carcinoma were included in this prospective study. All patients received one cycle TP(nedaplatin/paclitaxel) neoadjuvant chemotherapy and underwent conventional MRI and ¹⁸F-FDG PET/CT scans before and after treatment. All patients had radical hysterectomy after neoadjuvant chemotherapy. Based on MRI T2WI image, the chemotherapeutic responses were evaluated according to the RECIST 1.1. Based on PET/CT image, with the assistance of PET VCAR software, the chemotherapeutic responses were evaluated according to the EORTC criteria and PERCIST 1.0. We use the pathological five-tier Mandard TRG (tumor regression grade, TRG) as gold standard (TRG 1: No residual cancer cells; TRG 2 : Rare cancer cells ;TRG 3: Fibrosis outgrowing residual cancer; TRG 4 : Residual cancer outgrowing fibrosis; TRG 5: Absence of regressive change. The pre- and post- chemotherapy PET/CT scan were perform with similar dose of ¹⁸F-FDG and interval time. Concordance among these protocols was assessed using Cohen’s κ coefficient .

Results: Twenty of 21 patients have squamous carcinoma and only one patient has adenocarinoma. The FIGO stage of 21 patients range from IB1-IIB. Seven patients have lymph node metastasis, the other 14 patients have no metastasis. According to RECIST 1.1 criteria, one patient was evaluated as CR(complete response), 7 patients as PR(partial response) and 13 patients as SD(stable disease). Based on PERCIST 1.0 criteria, CMR (complete metabolic response) has 6 patients, PMR(partial metabolic response) has 11, and SMD(stable metabolic disease) has 4 patients. Sixteen patients were classified into PR, and only 5 patients into SD on the basis of EORTC criteria. In the light of TRG criteria, TRG 2 has only one patient; TRG 3 has 4 patents; TRG4 has 10 patents; TRG 5 has 6 patients. RECIST 1.1 and TRG classifications were discordant in 10 patients (κ=0.12, P > 0.05). EORTC and TRG response classifications were inconsistent with 4 patients(κ=0.54, P <0.05) , and ), and PERCIST 1.0 and TRG classifications were discordant in only 5 patients ( κ=0.40, P <0.05). Between RECIST 1.1 and EORTC response classifications, there were discrepancy in 8 patients (κ=0.32, P <0.05), and RECIST 1.1 and PERCIST 1.0 classifications in 9 patients (κ=0.25, P > 0.05)). EORTC and PERCIST 1.0 classifications were discordant in only 5 patients, resulting in better concordance (κ=0.30, P < 0.05).

Conclusion: EORTC criteria and PERCIST 1.0 are more sensitive and accurate than RECIST 1.1 for the detection of an early therapeutic response to chemotherapy in patients with cervical cancer. Research Support:This work was supported by a Grant for project research(81671771)from National Science Foundation of China and a grant (2016RAXYJ066) from Harbin science and technology bureau.