Objective:To explore the prognostic value of a combined model of baseline 18F-FDG PET/CT tumor metabolic parameters and clinical factors for predicting progression-free survival (PFS) in Hodgkin′s lymphoma (HL). Methods:From January 2014 to May 2023, 171 HL patients (102 males, 69 females; median age 40 years) who underwent 18F-FDG PET/CT before treatment at the First Affiliated Hospital of Nanjing Medical University and Nanjing Drum Tower Hospital were retrospectively collected. HL patients from the First Affiliated Hospital of Nanjing Medical University were classified as the training set (101 patients) and HL patients from Nanjing Drum Tower Hospital were classified as the validation set (70 patients). Clinical factors and tumor metabolic parameters associated with PFS were determined by multivariate Cox regression analysis, and then the combined model and the independent model of each factor were constructed respectively. The consistency index (C-index) and AUC were used to evaluate the predictive efficacy of models, and nomogram was constructed based on the optimal model, and calibration curves were used to assess the goodness of fit of the models. The differences in Kaplan-Meier survival curves of the high-risk and low-risk groups were compared using log-rank test. Results:The multivariate Cox regression analysis indicated that the independent prognostic factors associated with PFS were the Lugano staging (hazard ratio ( HR)=3.10, 95% CI: 1.17-8.23, P=0.023), total metabolic tumor volume (TMTV) ( HR=2.65, 95% CI: 1.23-5.74, P=0.014), and maximum distance between tumors ( Dmax) ( HR=2.23, 95% CI: 1.02-4.85, P=0.044). These factors were used to construct the combined model, with the highest prognostic efficacy of the C-index for the training and validation sets of 0.692 and 0.653, and the AUC of 0.732 and 0.697, respectively. The calibration curves demonstrated that the predictions made by the combined model were in high agreement with the actual results in both the training and validation sets. The Kaplan-Meier analysis revealed a significantly lower PFS rate in the high-risk group compared to the low-risk group both in training and validation sets ( χ2 values: 5.88 and 4.52, P values: 0.015 and 0.033). Conclusion:The combined model incorporating tumor metabolic parameters and clinical factors improves prognostic efficacy in predicting PFS in HL patients.

Objective:To explore the value of baseline 18F-FDG PET/CT metabolic parameters and related clinical factors in the prognostic assessment of patients with angioimmunoblastic T-cell lymphoma (AITL). Methods:From July 2013 to December 2023, 70 patients with AITL (44 males, 26 females, age (63.9±9.6) years) from Nanjing Drum Tower Hospital Clinical College of Nanjing Medical University (32 cases) as well as the First Affiliated Hospital of Nanjing Medical University (38 cases) who were diagnosed pathologically and underwent PET/CT imaging prior to treatment were retrospectively analyzed. PET/CT metabolic parameters (calculated using the 41%SUV max threshold method) and related clinical factors were collected. The optimal cut-off values of metabolic parameters were determined by using the ROC curve analysis. Cox proportional risk regression models were used for prognostic analyses, prediction models were constructed and efficacies were assessed by calibration curves and time-dependent ROC curves. Results:With the follow-up of 19.0(10.0, 33.3) months, disease progression or recurrence occurred in 51 patients, and 28 patients died. ROC curves showed that the optimal cut-off values on diagnosing AITL of total metabolic tumor volume (TMTV), total lesion glycolysis (TLG), and SUV max were 767.1cm 3, 2159.6g and 13.0, respectively. TMTV (hazard ratio ( HR)=0.485, 95% CI: 0.252-0.935, P=0.031) and gender ( HR=0.441, 95% CI: 0.236-0.824, P=0.010) were independent risk factors for progression-free survival (PFS); TMTV ( HR=0.422, 95% CI: 0.178-0.997, P=0.049) and treatment regimen ( HR=0.346, 95% CI: 0.154-0.777, P=0.010) were independent risk factors for overall survival (OS). Time-dependent ROC curves indicated that the combined model of TMTV combining gender or treatment regimen had better prognostic results in predicting PFS (AUCs: 0.67-0.82) or OS (AUCs: 0.62-0.80) in patients with AITL. The calibration curve showed the predicted values of the combined models were in good consistency with the actual values. Conclusions:The metabolic parameter TMTV is an independent risk factor for PFS and OS in patients with AITL. The combined model of TMTV combining gender or treatment regimen can effectively improve the prognostic prediction efficacy of PFS or OS in patients with AITL.

Objective:To investigate the value of a cross-combination machine learning approach in constructing a PET radiomics score (RadScore) for predicting early treatment response and prognosis in patients with primary gastrointestinal diffuse large B-cell lymphoma (PGI-DLBCL).Methods:This retrospective cohort study was conducted on 108 patients (59 males and 49 females, age (55.6±12.1) years) diagnosed with PGI-DLBCL between November 2016 and December 2021 at Nanjing Drum Tower Hospital, Affiliated Hospital of Medical School, Nanjing University ( n=85) and West China Hospital, Sichuan University ( n=23). Patients were divided into a training set ( n=86) and a validation set ( n=22) with the ratio of 8∶2 using stratified random sampling method. Seven machine learning models were employed to generate 49 feature selection-classification candidates, and the optimal candidate was selected to construct the RadScore, with five-fold cross-validation applied to determine the best-performing model. Logistic regression analysis was performed to identify risk factors for early treatment response, and a radiomics nomogram was developed by integrating RadScore with clinical predictors. Survival results between different groups of RadScore was compared by log-rank test. Results:Nineteen predictive features were selected from 111 radiomic features to construct the RadScore. In the training set, lactate dehydrogenase (LDH) (odds ratio ( OR)=3.53, 95% CI: 1.21-10.31, P=0.021), intestinal involvement ( OR=3.04, 95% CI: 1.04-8.88, P=0.042), total lesion glycolysis (TLG; OR=6.73, 95% CI: 2.23-20.29, P<0.001) and RadScore ( OR=15.11, 95% CI: 3.95-57.80, P<0.001) were identified as independent risk factors for predicting early treatment response. The combined model integrating RadScore, LDH, intestinal involvement, and TLG demonstrated good discriminatory ability for early treatment response (AUC=0.860 in the training set; AUC=0.902 in the validation set). Significant differences were observed in progression-free survival (PFS) and overall survival (OS) between different RadScore groups ( χ2 values: 13.92 and 8.56, both P<0.01). Conclusions:The machine learning-based RadScore may effectively predict survival outcomes in patients with PGI-DLBCL. The combined model integrating RadScore, clinical factors, and metabolic indicators can predict early treatment response in PGI-DLBCL patients.

Esophageal cancer(EC)is most commonly diagnosed in our country as an advanced esophageal cancer.With the continuous exploration and innovation of treatment methods,the comprehensive treatment scheme based on surgery is gradually improved.Immunotherapy is a new way to treat tumors in recent years.How to make better use of it in neoadjuvant therapy to create a better prognosis for EC is worth discussing.

Esophageal cancer(EC)is most commonly diagnosed in our country as an advanced esophageal cancer.With the continuous exploration and innovation of treatment methods,the comprehensive treatment scheme based on surgery is gradually improved.Immunotherapy is a new way to treat tumors in recent years.How to make better use of it in neoadjuvant therapy to create a better prognosis for EC is worth discussing.

Objective:To explore the prognostic value of a combined model of baseline 18F-FDG PET/CT tumor metabolic parameters and clinical factors for predicting progression-free survival (PFS) in Hodgkin′s lymphoma (HL). Methods:From January 2014 to May 2023, 171 HL patients (102 males, 69 females; median age 40 years) who underwent 18F-FDG PET/CT before treatment at the First Affiliated Hospital of Nanjing Medical University and Nanjing Drum Tower Hospital were retrospectively collected. HL patients from the First Affiliated Hospital of Nanjing Medical University were classified as the training set (101 patients) and HL patients from Nanjing Drum Tower Hospital were classified as the validation set (70 patients). Clinical factors and tumor metabolic parameters associated with PFS were determined by multivariate Cox regression analysis, and then the combined model and the independent model of each factor were constructed respectively. The consistency index (C-index) and AUC were used to evaluate the predictive efficacy of models, and nomogram was constructed based on the optimal model, and calibration curves were used to assess the goodness of fit of the models. The differences in Kaplan-Meier survival curves of the high-risk and low-risk groups were compared using log-rank test. Results:The multivariate Cox regression analysis indicated that the independent prognostic factors associated with PFS were the Lugano staging (hazard ratio ( HR)=3.10, 95% CI: 1.17-8.23, P=0.023), total metabolic tumor volume (TMTV) ( HR=2.65, 95% CI: 1.23-5.74, P=0.014), and maximum distance between tumors ( Dmax) ( HR=2.23, 95% CI: 1.02-4.85, P=0.044). These factors were used to construct the combined model, with the highest prognostic efficacy of the C-index for the training and validation sets of 0.692 and 0.653, and the AUC of 0.732 and 0.697, respectively. The calibration curves demonstrated that the predictions made by the combined model were in high agreement with the actual results in both the training and validation sets. The Kaplan-Meier analysis revealed a significantly lower PFS rate in the high-risk group compared to the low-risk group both in training and validation sets ( χ2 values: 5.88 and 4.52, P values: 0.015 and 0.033). Conclusion:The combined model incorporating tumor metabolic parameters and clinical factors improves prognostic efficacy in predicting PFS in HL patients.

Objective:To explore the value of baseline 18F-FDG PET/CT metabolic parameters and related clinical factors in the prognostic assessment of patients with angioimmunoblastic T-cell lymphoma (AITL). Methods:From July 2013 to December 2023, 70 patients with AITL (44 males, 26 females, age (63.9±9.6) years) from Nanjing Drum Tower Hospital Clinical College of Nanjing Medical University (32 cases) as well as the First Affiliated Hospital of Nanjing Medical University (38 cases) who were diagnosed pathologically and underwent PET/CT imaging prior to treatment were retrospectively analyzed. PET/CT metabolic parameters (calculated using the 41%SUV max threshold method) and related clinical factors were collected. The optimal cut-off values of metabolic parameters were determined by using the ROC curve analysis. Cox proportional risk regression models were used for prognostic analyses, prediction models were constructed and efficacies were assessed by calibration curves and time-dependent ROC curves. Results:With the follow-up of 19.0(10.0, 33.3) months, disease progression or recurrence occurred in 51 patients, and 28 patients died. ROC curves showed that the optimal cut-off values on diagnosing AITL of total metabolic tumor volume (TMTV), total lesion glycolysis (TLG), and SUV max were 767.1cm 3, 2159.6g and 13.0, respectively. TMTV (hazard ratio ( HR)=0.485, 95% CI: 0.252-0.935, P=0.031) and gender ( HR=0.441, 95% CI: 0.236-0.824, P=0.010) were independent risk factors for progression-free survival (PFS); TMTV ( HR=0.422, 95% CI: 0.178-0.997, P=0.049) and treatment regimen ( HR=0.346, 95% CI: 0.154-0.777, P=0.010) were independent risk factors for overall survival (OS). Time-dependent ROC curves indicated that the combined model of TMTV combining gender or treatment regimen had better prognostic results in predicting PFS (AUCs: 0.67-0.82) or OS (AUCs: 0.62-0.80) in patients with AITL. The calibration curve showed the predicted values of the combined models were in good consistency with the actual values. Conclusions:The metabolic parameter TMTV is an independent risk factor for PFS and OS in patients with AITL. The combined model of TMTV combining gender or treatment regimen can effectively improve the prognostic prediction efficacy of PFS or OS in patients with AITL.

Objective:To investigate the value of a cross-combination machine learning approach in constructing a PET radiomics score (RadScore) for predicting early treatment response and prognosis in patients with primary gastrointestinal diffuse large B-cell lymphoma (PGI-DLBCL).Methods:This retrospective cohort study was conducted on 108 patients (59 males and 49 females, age (55.6±12.1) years) diagnosed with PGI-DLBCL between November 2016 and December 2021 at Nanjing Drum Tower Hospital, Affiliated Hospital of Medical School, Nanjing University ( n=85) and West China Hospital, Sichuan University ( n=23). Patients were divided into a training set ( n=86) and a validation set ( n=22) with the ratio of 8∶2 using stratified random sampling method. Seven machine learning models were employed to generate 49 feature selection-classification candidates, and the optimal candidate was selected to construct the RadScore, with five-fold cross-validation applied to determine the best-performing model. Logistic regression analysis was performed to identify risk factors for early treatment response, and a radiomics nomogram was developed by integrating RadScore with clinical predictors. Survival results between different groups of RadScore was compared by log-rank test. Results:Nineteen predictive features were selected from 111 radiomic features to construct the RadScore. In the training set, lactate dehydrogenase (LDH) (odds ratio ( OR)=3.53, 95% CI: 1.21-10.31, P=0.021), intestinal involvement ( OR=3.04, 95% CI: 1.04-8.88, P=0.042), total lesion glycolysis (TLG; OR=6.73, 95% CI: 2.23-20.29, P<0.001) and RadScore ( OR=15.11, 95% CI: 3.95-57.80, P<0.001) were identified as independent risk factors for predicting early treatment response. The combined model integrating RadScore, LDH, intestinal involvement, and TLG demonstrated good discriminatory ability for early treatment response (AUC=0.860 in the training set; AUC=0.902 in the validation set). Significant differences were observed in progression-free survival (PFS) and overall survival (OS) between different RadScore groups ( χ2 values: 13.92 and 8.56, both P<0.01). Conclusions:The machine learning-based RadScore may effectively predict survival outcomes in patients with PGI-DLBCL. The combined model integrating RadScore, clinical factors, and metabolic indicators can predict early treatment response in PGI-DLBCL patients.

Objective:To explore the value of baseline PET/CT parameters for predicting prognosis in patients with double-expression lymphoma (DEL).Methods:The clinical and 18F-FDG PET/CT data of 118 patients (66 males, 52 females; age: 28-85 years) with diffuse large B-cell lymphoma (DLBCL) diagnosed in Nanjing Drum Tower Hospital, Affiliated Hospital of Medical School, Nanjing University and the First Affiliated Hospital of Nanjing Medical University from June 2015 to September 2022 were retrospectively analyzed. The optimal thresholds for SUV max, total metabolic tumor volume (TMTV) and total lesion glycolysis (TLG) in predicting overall survival (OS) rate were determined using ROC curve analysis. Univariate and multivariate analyses, along with Kaplan-Meier survival analysis were performed to construct a survival prediction model. The effect of the model was evaluated by the calibration curve for the model, the time-dependent ROC curve analysis and decision curve analysis. Results:As of the last follow-up, 25 patients died, and the OS rate was 78.8%(93/118). The AUC of the ROC curve for TMTV was 0.705, with a corresponding optimal threshold of 230.9 cm 3. In multivariate analysis, Eastern Cooperative Oncology Group performance status (ECOG PS) score (hazard ratio ( HR)=3.886, 95% CI: 1.455-10.375; P=0.007) and TMTV ( HR=4.649, 95% CI: 1.665-12.979; P=0.003) were identified as independent predictors of OS. The combined model of ECOG PS score and TMTV was superior to ECOG PS score model and TMTV model alone in predicting OS. Conclusion:TMTV, a metabolic indicator, and ECOG PS score, a clinical risk factor, are independent predictors of OS in patients with DEL, and their combination can provide more accurate prognostic predictions.

Objective:To explore the prognostic value of 18F-FDG PET-based radiomics features by machine learning in older patients(≥60 years) with diffuse large B-cell lymphoma (DLBCL). Methods:A total of 166 older patients (88 males, 78 females, age: 60-93 years) with DLBCL who underwent pre-therapy 18F-FDG PET/CT from March 2011 to November 2019 were enrolled in the retrospective study. There were 115 patients in training cohort and 51 patients in validation cohort. The lesions in PET images were manually drawn and the obtained radiomics features from patients in training cohort were selected by the least absolute shrinkage and selection operator (LASSO), random forest (RF), and extreme gradient boosting (Xgboost), and then classified by support vector machine (SVM) to build radiomics signatures (RS) for predicting overall survival (OS). A multi-parameter model was constructed by using Cox proportional hazard model and assessed by concordance index (C-index). Results:A total of 1 421 PET radiomics features were extracted and 10 features were selected to build RS. The univariate Cox regression analysis showed that RS was a predictor of OS (hazard ratio ( HR)=5.685, 95% CI: 2.955-10.939; P<0.001). The multi-parameter model that incorporated RS, metabolic metrics, and clinical risk factors, exhibited significant prognostic superiority over the clinical model, PET-based model, and the National Comprehensive Cancer Network International Prognostic Index (NCCN-IPI) in terms of OS (training cohort: C-index: 0.752 vs 0.737 vs 0.739 vs 0.688; validation cohort: C-index: 0.845 vs 0.798 vs 0.844 vs 0.775). Conclusions:RS can be used as a survival predictor for older patients(≥60 years) with DLBCL. Furthermore, the multi-parameter model incorporating RS is able to successfully predict prognosis.