Objective  To analysis the value of four evaluation indicators based on interim ¹⁸F-fluorodeoxyglucose (FDG) PET/CT (iPET/CT) in predicting the response to primary chemotherapy and prognosis of patients with stage Ⅲ/Ⅳ extranodal diffuse large B-cell lymphoma (EN-DLBCL).

Methods  A retrospective analysis was conducted on the imaging and clinical data of 56 newly diagnosed patients (27 males and 29 females; aged 27−28 years old with a median age of 68.5) with stage Ⅲ/Ⅳ EN-DLBCL confirmed by histopathological examination or follow-up in Beijing Hospital from May 2011 to April 2020. Follow-up was performed through the hospital's electronic medical record system or phone call. The receiver operating characteristic (ROC) curve was used to select the optimal cut-off for the maximum standardized uptake value of iPET/CT lesions (iSUV_max), the reduction percentage of lesions' SUV_max on interim and baseline PET/CT (ΔSUV_max%), and the ratio of lesions' SUV_max to liver SUV_max (iLLR). Patients divided into groups using the optimal cutoff and the interim Deauville 5-point scale (DS). The predictive value of iPET/CT indicators on the efficacy and prognosis of primary chemotherapy in patients with stage Ⅲ/Ⅳ EN-DLBCL was analyzed. χ² test was employed to compare the count data. Kaplan-Meier method and Log-rank test were applied to analyze the patients' 2-year progression-free survival (PFS) rates, overall survival (OS) rates, and differences between groups. Univariate and multivariate Cox proportional risk regression analysis were utilized to analyze the predictive value of iSUV_max, ΔSUV_max%, iLLR, and DS for the efficacy and prognosis of primary chemotherapy in patients with stage Ⅲ/Ⅳ EN-DLBCL.

Results  At the end of primary chemotherapy, 31 patients (55.4%) achieved complete remission (CR). The CR rates in the iSUV_max<4.9, ΔSUV_max%≥87.0%, iLLR<1.31, and DS 1−3 groups were higher than those in the iSUV_max≥4.9, ΔSUV_max%<87.0%, iLLR≥1.31, and DS 4−5 groups (74.2% (23/31) vs. 32.0%(8/25), 76.2%(16/21) vs. 42.9%(15/35), 67.6% (23/34) vs. 36.4%(8/22), and 76.0%(19/25) vs. 38.7%(12/31)), the differences were statistically significantly (χ²=9.970, 5.901, 5.290, 7.787; all P<0.05). The follow-up time was 4−103 months with a median time of 22 months. At the end of the follow-up, 19 patients (33.9%) had progressed, relapsed, or died. The 2-year PFS rates in the iSUV_max<4.9, iLLR<1.63, and DS 1−3 groups (83.9%, 83.7%, and 84.0%) were significantly higher than those in the iSUV_max≥4.9, iLLR≥1.63, and DS 4−5 groups (50.4%, 37.6%, and 57.5%), the differences were statistically significantly (χ²=6.117, 10.478, 4.116; all P<0.05). The 2-year OS rate in the iLLR<1.63 group (90.3%) was significantly higher than that in the iLLR≥1.63 group (67.4%), the difference was statistically significantly (χ²=5.220, P<0.05). Univariate Cox proportional risk regression analysis showed that iSUV_max and iLLR were predict factors for patients' PFS time (HR=2.883(95%CI: 1.124−7.398), 3.344(95%CI: 1.339−8.353); both P<0.05), and iLLR was an independent predict factor of patients' OS time (HR=4.445(95%CI: 1.266−15.607), P<0.05). Furthermore, multivariate Cox proportional risk regression analysis confirmed that iLLR was the independent predict factor of PFS and OS time in patients with stage Ⅲ/Ⅳ EN-DLBCL (HR=3.344(95%CI: 1.339−8.353), 4.445(95%CI: 1.266−15.607); both P<0.050).

Conclusions  The four evaluation indicators of iPET/CT in patients with stage Ⅲ/Ⅳ EN-DLBCL can predict whether CR can be achieved at the end of primary chemotherapy. iLLR can be considered as an independent predict factor of patients' PFS and OS time.