Radiomic Features of 18F-FDG PET in Hodgkin Lymphoma Are Predictive of Outcomes

Challenges and limitations in applying radiomics to PET imaging: Possible opportunities and avenues for research

Radiomics, the high-throughput extraction of quantitative imaging features from medical images, holds immense potential for advancing precision medicine in oncology and beyond. While radiomics applied to positron emission tomography (PET) imaging offers unique insights into tumor biology and treatment response, it is imperative to elucidate the challenges and constraints inherent in this domain to facilitate their translation into clinical practice. This review examines the challenges and limitations of applying radiomics to PET imaging, synthesizing findings from the last five years (2019-2023) and highlights the significance of addressing these challenges to realize the full clinical potential of radiomics in oncology and molecular imaging. A comprehensive search was conducted across multiple electronic databases, including PubMed, Scopus, and Web of Science, using keywords relevant to radiomics issues in PET imaging. Only studies published in peer-reviewed journals were eligible for inclusion in this review. Although many studies have highlighted the potential of radiomics in predicting treatment response, assessing tumor heterogeneity, enabling risk stratification, and personalized therapy selection, various challenges regarding the practical implementation of the proposed models still need to be addressed. This review illustrates the challenges and limitations of radiomics in PET imaging across various cancer types, encompassing both phantom and clinical investigations. The analyzed studies highlight the importance of reproducible segmentation methods, standardized pre-processing and post-processing methodologies, and the need to create large multicenter studies registered in a centralized database to promote the continuous validation and clinical integration of radiomics into PET imaging.

Prognostic value of the combination of volume, massiveness and fragmentation parameters measured on baseline FDG pet in high-burden follicular lymphoma

The prognostic value of radiomic quantitative features measured on pre-treatment 18F-FDG PET/CT was investigated in patients with follicular lymphoma (FL). We conducted a retrospective study of 126 FL patients (grade 1-3a) diagnosed between 2006 and 2020. A dozen of PET/CT-derived features were extracted via a software (Oncometer3D) from baseline 18F-FDG PET/CT images. The receiver operating characteristic (ROC) curve, Kaplan-Meier method and Cox analysis were used to assess the prognostic factors for progression of disease within 24 months (POD24) and progression-free survival at 24 months. Four different clusters were identified among the twelve PET parameters analyzed: activity, tumor burden, fragmentation-massiveness and dispersion. On ROC analyses, TMTV, the total metabolic tumor volume, had the highest AUC (0.734) followed by medPCD, the median distance between the centroid of the tumors and their periphery (AUC: 0.733). Patients with high TMTV (HR = 4.341; p < 0.001), high Tumor Volume Surface Ratio (TVSR) (HR = 3.204; p < 0.003) and high medPCD (HR = 4.507; p < 0.001) had significantly worse prognosis in both Kaplan-Meier and Cox univariate analyses. Furthermore, a synergistic effect was observed in Kaplan-Meier and Cox analyses combining these three PET/CT-derived parameters (HR = 12.562; p < 0.001). Having two or three high parameters among TMTV, TVSR and medPCD was able to predict POD24 status with a specificity of 68% and a sensitivity of 75%. TMTV, TVSR and baseline medPCD are strong prognostic factors in FL and their combination better predicts disease prognosis.

A different perspective on PET/CT before treatment in patients with Hodgkin lymphoma: importance of volumetric and dissemination parameters

The aim of this study is to investigate the role of the combination of volumetric and dissemination parameters obtained from pretreatment 18-fluorodeoxyglucose positron emission tomography/computed tomography (18F-FDG PET/CT) in predicting the interim response and progression status in patients with Hodgkin lymphoma (HL). Pretreatment PET/CT images of HL patients were analyzed with LIFEx software, and volumes of interest (VOIs) were drawn with a fixed SUV 4.0 threshold. MTV, SUVmax, and TLG values were obtained from each VOI. Total MTV (tMTV) was calculated by summing the MTV values in all VOIs, and similarly, total TLG (tTLG) was obtained by summing the TLG values. The distance between the centers of the lesions was noted as Dmax, and the distance between the outermost voxels of the lesions as DmaxVox. tMTV/DmaxVox was calculated by dividing the tMTV value by the DmaxVox value, and tTLG/DmaxVox was calculated by dividing the tTLG value by the DmaxVox value. The correlation of pretreatment PET parameters with response groups (complete/poor) and relapse/progression status (stable/progressive) was statistically evaluated. A total of 52 patients were included in the study. Bulky disease, tMTV, tTLG, and tMTV/DmaxVox values were found to be significantly higher in the poor response group. tMTV > 190.60 ml was found to be the only prognostic factor predicting interim PET response. The tMTV/DmaxVox and tTLG/DmaxVox showed statistically significant differences between the groups with and without progression. tMTV/DmaxVox > 7.70 was found to be the only prognostic factor in predicting relapse/progression. The evaluation of tumor burden and dissemination together in 18F-FDG PET/CT before treatment in patients with HL can help us to predict the results of the patients.

Predictive value of Dmax and %ΔSUVmax of 18F-FDG PET/CT for the prognosis of patients with diffuse large B-cell lymphoma

PURPOSE

To investigate the prognosis value of a combined model based on 18F-fluoro-deoxyglucose positron emission tomography-computed tomography (18F-FDG PET-CT) baseline and interim parameters in patients with diffuse large B-cell lymphoma (DLBCL).

METHODS

We retrospectively analyzed the PET metabolic parameters and clinical data of 154 DLBCL patients between December 2015 and October 2020. All of these patients underwent 18F-FDG PET/CT scan before treatment and after three or four courses of chemotherapy. The optimal cut-off values for quantitative variables were determined by the receiver operating characteristic (ROC) curve. The baseline and interim PET/CT parameters, which respectively included maximum standardized uptake value (SUVmax0), total metabolic tumor volume (TMTV0), standardized total metabolic tumor volume (STMTV0), and the distance between the two furthest lesions (Dmax) and total tumor lesion glycolysis (TTLG1), SUVmax1, TMTV1, and the rate of change of SUVmax (%ΔSUVmax), and clinical characteristics were analyzed by chi-squared test, Kaplan-Meier survival curve, and Cox regression analysis.

RESULTS

Of 154 patients, 35 exhibited disease progression or recurrence. ROC analysis revealed that baseline 18F-FDG PET/CT metabolic parameters, including maximum standardized uptake value (SUVmax0), total metabolic tumor volume (TMTV0), standardized total metabolic tumor volume (STMTV0), and the distance between the two furthest lesions (Dmax), along with interim 18F-FDG PET/CT metabolic parameters such as total tumor lesion glycolysis (TTLG1), SUVmax1, TMTV1, and the rate of change of SUVmax (%ΔSUVmax), were predictive of relapse or progression in DLBCL patients (P < 0.05). The chi-squared test showed that TMTV0, STMTV0, Dmax, SUVmax1, TMTV1, TTLG1, %ΔSUVmax, Deauville score, IPI, Ann Arbor stage, and LDH were associated with patient prognosis (P < 0.05). Multivariate Cox regression analysis showed that Dmax (P = 0.021) and %ΔSUVmax (P = 0.030) were independent predictors of prognosis in DLBCL patients. There were statistically significant differences in PFS among the three groups with high, intermediate, and low risk according to the combination model (P < 0.001). The combination model presented higher predictive efficacy than single indicators.

CONCLUSION

The combined model of baseline parameter Dmax and intermediate parameter %ΔSUVmax of 18F-FDG PET/CT improved the predictive efficacy of PFS and contributed to the risk stratification of patients, providing a reference for clinical individualization and precision treatment.

18F-FDG PET/CT Maximum Tumor Dissemination (Dmax) in Lymphoma: A New Prognostic Factor?

Recently, several studies introduced the potential prognostic usefulness of maximum tumor dissemination (Dmax) measured by 2-deoxy-2-fluorine-18-fluoro-D-glucose positron-emission tomography/computed tomography (¹⁸F-FDG PET/CT). Dmax is a simple three-dimensional feature that represents the maximal distance between the two farthest hypermetabolic PET lesions. A comprehensive computer literature search of PubMed/MEDLINE, Embase, and Cochrane libraries was conducted, including articles indexed up to 28 February 2023. Ultimately, 19 studies analyzing the value of ¹⁸F-FDG PET/CT Dmax in patients with lymphomas were included. Despite their heterogeneity, most studies showed a significant prognostic role of Dmax in predicting progression-free survival (PFS) and overall survival (OS). Some articles showed that the combination of Dmax with other metabolic features, such as MTV and interim PET response, proved to better stratify the risk of relapse or death. However, some methodological open questions need to be clarified before introducing Dmax into clinical practice.

Combination of FDG PET/CT Radiomics and Clinical Parameters for Outcome Prediction in Patients with Hodgkin’s Lymphoma

Purpose: The aim of the study is to evaluate the prognostic value of a joint evaluation of PET and CT radiomics combined with standard clinical parameters in patients with HL. Methods: Overall, 88 patients (42 female and 46 male) with a median age of 43.3 (range 21–85 years) were included. Textural analysis of the PET/CT images was performed using freely available software (LIFE X). 65 radiomic features (RF) were evaluated. Univariate and multivariate models were used to determine the value of clinical characteristics and FDG PET/CT radiomics in outcome prediction. In addition, a binary logistic regression model was used to determine potential predictors for radiotherapy treatment and odds ratios (OR), with 95% confidence intervals (CI) reported. Features relevant to survival outcomes were assessed using Cox proportional hazards to calculate hazard ratios with 95% CI. Results: albumin (p = 0.034) + ALP (p = 0.028) + CT radiomic feature GLRLM GLNU mean (p = 0.012) (Area under the curve (AUC): 95% CI (86.9; 100.0)—Brier score: 3.9, 95% CI (0.1; 7.8) remained significant independent predictors for PFS outcome. PET-SHAPE Sphericity (p = 0.033); CT grey-level zone length matrix with high gray-level zone emphasis (GLZLM SZHGE mean (p = 0.028)); PARAMS XSpatial Resampling (p = 0.0091) as well as hemoglobin results (p = 0.016) remained as independent factors in the final model for a binary outcome as predictors of the need for radiotherapy (AUC = 0.79). Conclusion: We evaluated the value of baseline clinical parameters as well as combined PET and CT radiomics in HL patients for survival and the prediction of the need for radiotherapy treatment. We found that different combinations of all three factors/features were independently predictive of the here evaluated endpoints.

The added prognostic values of baseline PET dissemination parameter in patients with angioimmunoblastic T-cell lymphoma

To explore the prognostic values of baseline 2-deoxy-2-[18F] fluoro-D-glucose (FDG) positron emission tomography/computed tomography (PET/CT) dissemination parameter in angioimmunoblastic T-cell lymphoma (AITL) and its added values to total metabolic tumour volume (TMTV). Eighty-one AITL patients with at least two FDG-avid lesions in baseline PET/CT were retrospectively included. PET parameters including TMTV and the distance between the two lesions that are the furthest apart (Dmax) were obtained. Univariate Cox analysis showed that both Dmax and TMTV were risk factors for progression-free survival (PFS) and overall survival (OS). Multivariate Cox analysis models of different combinations showed that high Dmax (> 65.7 cm) could independently predict both PFS and OS, while high TMTV (>456.6 cm3) was only significant for OS. A concise PET model based on TMTV and Dmax can effectively risk-stratify patients. PFS and OS rates were significantly lower in patients with high Dmax and high TMTV than in patients with low Dmax and low TMTV (3-year PFS rate: 15.0% vs. 48.7%, p = 0.001; 3-year OS rate: 27.6% vs. 79.0%, p < 0.001). Dmax can directly reflect the disease dissemination characteristic and has a significant prognostic value for FDG-avid AITL patients. It has the potential to be introduced into new risk stratification models for tailored treatment.

Development and validation of nomograms by radiomic features on ultrasound imaging for predicting overall survival in patients with primary nodal diffuse large B-cell lymphoma

Objectives

To develop and validate a nomogram to predict the overall survival (OS) of patients with primary nodal diffuse large B-cell lymphoma(N-DLBCL) based on radiomic features and clinical features.

Materials and methods

A retrospective analysis was performed on 145 patients confirmed with N-DLBCL and they were randomly assigned to training set(n=78), internal validation set(n=33), external validation set(n=34). First, a clinical model (model 1) was established according to clinical features and ultrasound (US) results. Then, based on the radiomics features extracted from conventional ultrasound images, a radiomic signature was constructed (model 2), and the radiomics score (Rad-Score) was calculated. Finally, a comprehensive model was established (model 3) combined with Rad-score and clinical features. Receiver operating characteristic (ROC) curves were employed to evaluate the performance of model 1, model 2 and model 3. Based on model 3, we plotted a nomogram. Calibration curves were used to test the effectiveness of the nomogram, and decision curve analysis (DCA) was used to asset the nomogram in clinical use.

Results

According to multivariate analysis, 3 clinical features and Rad-score were finally selected to construct the model 3, which showed better predictive value for OS in patients with N-DLBCL than mode 1 and model 2 in training (AUC,0. 891 vs. 0.779 vs.0.756), internal validation (AUC, 0.868 vs. 0.713, vs.0.756) and external validation (AUC, 914 vs. 0.866, vs.0.789) sets. Decision curve analysis demonstrated that the nomogram based on model 3 was more clinically useful than the other two models.

Conclusion

The developed nomogram is a useful tool for precisely analyzing the prognosis of N-DLBCL patients, which could help clinicians in making personalized survival predictions and assessing individualized clinical options.

Utility of pre-treatment FDG PET/CT-derived machine learning models for outcome prediction in classical Hodgkin lymphoma

Objectives

Relapse occurs in ~20% of patients with classical Hodgkin lymphoma (cHL) despite treatment adaption based on 2-deoxy-2-[¹⁸F]fluoro-d-glucose positron emission tomography/computed tomography response. The objective was to evaluate pre-treatment FDG PET/CT–derived machine learning (ML) models for predicting outcome in patients with cHL.

Methods

All cHL patients undergoing pre-treatment PET/CT at our institution between 2008 and 2018 were retrospectively identified. A 1.5 × mean liver standardised uptake value (SUV) and a fixed 4.0 SUV threshold were used to segment PET/CT data. Feature extraction was performed using PyRadiomics with ComBat harmonisation. Training (80%) and test (20%) cohorts stratified around 2-year event-free survival (EFS), age, sex, ethnicity and disease stage were defined. Seven ML models were trained and hyperparameters tuned using stratified 5-fold cross-validation. Area under the curve (AUC) from receiver operator characteristic analysis was used to assess performance.

Results

A total of 289 patients (153 males), median age 36 (range 16–88 years), were included. There was no significant difference between training (n = 231) and test cohorts (n = 58) (p value > 0.05). A ridge regression model using a 1.5 × mean liver SUV segmentation had the highest performance, with mean training, validation and test AUCs of 0.82 ± 0.002, 0.79 ± 0.01 and 0.81 ± 0.12. However, there was no significant difference between a logistic model derived from metabolic tumour volume and clinical features or the highest performing radiomic model.

Conclusions

Outcome prediction using pre-treatment FDG PET/CT–derived ML models is feasible in cHL patients. Further work is needed to determine optimum predictive thresholds for clinical use.

Key points

• A fixed threshold segmentation method led to more robust radiomic features.

• A radiomic-based model for predicting 2-year event-free survival in classical Hodgkin lymphoma patients is feasible.

• A predictive model based on ridge regression was the best performing model on our dataset.

Supplementary Information

The online version contains supplementary material available at 10.1007/s00330-022-09039-0.

Prediction of prognosis and pathologic grade in follicular lymphoma using 18F-FDG PET/CT

Purpose

We investigated the utility of a new baseline PET parameter expressing lesion dissemination and metabolic parameters for predicting progression-free survival (PFS) and pathologic grade in follicular lymphoma (FL).

Methods

The baseline ¹⁸F-FDG PET/CT images of 126 patients with grade 1–3A FL were retrospectively analyzed. A novel PET/CT parameter characterizing lesion dissemination, the distance between two lesions that were furthest apart (D_max), was calculated. The total metabolic tumor volume and total lesion glycolysis (TLG) were computed by using 41% of the maximum standardized uptake value (SUV_max) thresholding method.

Results

The 5-year PFS rate was 51.9% for all patients. In the multivariate analysis, high D_max [P = 0.046; hazard ratio (HR) = 2.877], high TLG (P = 0.004; HR = 3.612), and elevated serum lactate dehydrogenase (P = 0.041; HR = 2.287) were independent predictors of PFS. A scoring system for prognostic stratification was established based on these three adverse factors, and the patients were classified into three risk categories: low risk (zero to one factor, n = 75), intermediate risk (two adverse factors, n = 29), and high risk (three adverse factors, n = 22). Patients in the high-risk group had a shorter 3-year PFS (21.7%) than those in the low- and intermediate-risk groups (90.6 and 44.6%, respectively) (P < 0.001). The C-index of our scoring system for PFS (0.785) was superior to the predictive capability of the Follicular Lymphoma International Prognostic Index (FLIPI), FLIPI2, and PRIMA-Prognostic Index (C-index: 0.628–0.701). The receiver operating characteristic curves and decision curve analysis demonstrated that the scoring system had better differentiation and clinical utility than these existing indices. In addition, the median SUV_max was significantly higher in grade 3A (36 cases) than in grades 1 and 2 FL (90 cases) (median: 13.63 vs. 11.45, P = 0.013), but a substantial overlap existed (range: 2.25–39.62 vs. 3.17–39.80).

Conclusion

TLG and D_max represent two complementary aspects of the disease, capturing the tumor burden and lesion dissemination. TLG and D_max are promising metrics for identifying patients at a high risk of progression or relapse. Additionally, SUV_max seems to have some value for distinguishing grade 3A from low-grade FL but cannot substitute for biopsy.

Radiomics in Oncological PET Imaging: A Systematic Review-Part 2, Infradiaphragmatic Cancers, Blood Malignancies, Melanoma and Musculoskeletal Cancers

The objective of this review was to summarize published radiomics studies dealing with infradiaphragmatic cancers, blood malignancies, melanoma, and musculoskeletal cancers, and assess their quality. PubMed database was searched from January 1990 to February 2022 for articles performing radiomics on PET imaging of at least 1 specified tumor type. Exclusion criteria includd: non-oncological studies; supradiaphragmatic tumors; reviews, comments, cases reports; phantom or animal studies; technical articles without a clinically oriented question; studies including <30 patients in the training cohort. The review database contained PMID, first author, year of publication, cancer type, number of patients, study design, independent validation cohort and objective. This database was completed twice by the same person; discrepant results were resolved by a third reading of the articles. A total of 162 studies met inclusion criteria; 61 (37.7%) studies included >100 patients, 13 (8.0%) were prospective and 61 (37.7%) used an independent validation set. The most represented cancers were esophagus, lymphoma, and cervical cancer (n = 24, n = 24 and n = 19 articles, respectively). Most studies focused on 18F-FDG, and prognostic and response to treatment objectives. Although radiomics and artificial intelligence are technically challenging, new contributions and guidelines help improving research quality over the years and pave the way toward personalized medicine.

Tumor Size Measurements for Predicting Hodgkin’s and Non-Hodgkin’s Lymphoma Response to Treatment

The purpose of this study was to investigate the value of tumor size measurements as prognostic indicators of treatment outcome of Hodgkin’s and Non-Hodgkin’s lymphomas. ¹⁸F-FDG PET/CT exams before and after treatment were analyzed and metabolic and anatomic parameters—tumor maximum diameter, tumor maximum area, tumor volume, and maximum standardized uptake value (SUVmax)—were determined manually by an expert and automatically by a computer algorithm on PET and CT images. Results showed that the computer algorithm measurements did not correlate well with the expert’s standard maximum tumor diameter measurements but yielded better three dimensional metrics that could have clinical value. SUVmax was the strongest prognostic indicator of the clinical outcome after treatment, followed by the automated metabolic tumor volume measurements and the expert’s metabolic maximum diameter measurements. Anatomic tumor measurements had poor prognostic value. Metabolic volume measurements, although promising, did not significantly surpass current standard of practice, but automated measurements offered a significant advantage in terms of time and effort and minimized biases and variances in the PET measurements. Overall, considering the limited value of tumor size in predicting response to treatment, a paradigm shift seems necessary in order to identify robust prognostic markers in PET/CT; radiomics, namely combinations of anatomy, metabolism, and imaging, may be an option.