Does FDG PET-Based Radiomics Have an Added Value for Prediction of Overall Survival in Non-Small Cell Lung Cancer?

Objectives: Radiomics and machine learning are innovative approaches to improve the clinical management of NSCLC. However, there is less information about the additive value of FDG PET-based radiomics compared with clinical and imaging variables. Methods: This retrospective study included 320 NSCLC patients who underwent PET/CT with FDG at initial staging. VOIs were placed on primary tumors only. We included a total of 94 variables, including 87 textural features extracted from PET studies, SUVmax, MTV, TLG, TNM stage, histology, age, and gender. We used the least absolute shrinkage and selection operator (LASSO) regression to select variables with the highest predictive value. Although several radiomics variables are available, the added value of these predictors compared with clinical and imaging variables is still under evaluation. Three hundred and twenty NSCLC patients were included in this retrospective study and underwent 18F-FDG PET/CT at initial staging. In this study, we evaluated 94 variables, including 87 textural features, SUVmax, MTV, TLG, TNM stage, histology, age, and gender. Image-based predictors were extracted from a volume of interest (VOI) positioned on the primary tumor. The least absolute shrinkage and selection operator (LASSO) Cox regression was used to reduce the number of variables and select only those with the highest predictive value. The predictive model implemented with the variables selected using the LASSO analysis was compared with a reference model using only a tumor stage and SUVmax. Results: NGTDM coarseness, SUVmax, and TNM stage survived the LASSO analysis and were used for the radiomic model. The AUCs obtained from the reference and radiomic models were 80.82 (95%CI, 69.01-92.63) and 81.02 (95%CI, 69.07-92.97), respectively (p = 0.98). The median OS in the reference model was 17.0 months in high-risk patients (95%CI, 11-21) and 113 months in low-risk patients (HR 7.47, p < 0.001). In the radiomic model, the median OS was 16.5 months (95%CI, 11-20) and 113 months in high- and low-risk groups, respectively (HR 9.64, p < 0.001). Conclusions: Our results indicate that a radiomic model composed using the tumor stage, SUVmax, and a selected radiomic feature (NGTDM_Coarseness) predicts survival in NSCLC patients similarly to a reference model composed only by the tumor stage and SUVmax. Replication of these preliminary results is necessary.

Multivariable Risk Modelling and Survival Analysis with Machine Learning in SARS-CoV-2 Infection.

AIM: To evaluate the performance of a machine learning model based on demographic variables, blood tests, pre-existing comorbidities, and computed tomography(CT)-based radiomic features to predict critical outcome in patients with acute respiratory syndrome coronavirus 2 (SARS-CoV-2). METHODS: We retrospectively enrolled 694 SARS-CoV-2-positive patients. Clinical and demographic data were extracted from clinical records. Radiomic data were extracted from CT. Patients were randomized to the training (80%, n = 556) or test (20%, n = 138) dataset. The training set was used to define the association between severity of disease and comorbidities, laboratory tests, demographic, and CT-based radiomic variables, and to implement a risk-prediction model. The model was evaluated using the C statistic and Brier scores. The test set was used to assess model prediction performance. RESULTS: Patients who died (n = 157) were predominantly male (66%) over the age of 50 with median (range) C-reactive protein (CRP) = 5 [1, 37] mg/dL, lactate dehydrogenase (LDH) = 494 [141, 3631] U/I, and D-dimer = 6.006 [168, 152.015] ng/mL. Surviving patients (n = 537) had median (range) CRP = 3 [0, 27] mg/dL, LDH = 484 [78, 3.745] U/I, and D-dimer = 1.133 [96, 55.660] ng/mL. The strongest risk factors were D-dimer, age, and cardiovascular disease. The model implemented using the variables identified using the LASSO Cox regression analysis classified 90% of non-survivors as high-risk individuals in the testing dataset. In this sample, the estimated median survival in the high-risk group was 9 days (95% CI; 9-37), while the low-risk group did not reach the median survival of 50% (p < 0.001). CONCLUSIONS: A machine learning model based on combined data available on the first days of hospitalization (demographics, CT-radiomics, comorbidities, and blood biomarkers), can identify SARS-CoV-2 patients at risk of serious illness and death.

Assessment of Response to Immunotherapy in Patients with Hodgkin Lymphoma: Towards Quantifying Changes in Tumor Burden Using FDG-PET/CT.

Immune checkpoint inhibitors are currently the standard of care for many advanced solid tumors, and they have been recently approved for the treatment of relapsed/refractory Hodgkin lymphoma and primary mediastinal B cell lymphoma. Assessments of the response to immunotherapy may be complicated by the occurrence of the flare/pseudoprogression phenomenon, consisting of initial tumor enlargement and even the appearance of new lesions, followed by a response, which may initially be indistinguishable from true progression. There have been efforts to characterize and capture the new patterns of response observed during immunotherapy, namely, pseudoprogression and delayed response, and several immune-related response criteria have been proposed. Confirming progression on a subsequent scan and measuring the total tumor burden are both common in immune-related criteria. Due to the peculiarity of hematologic malignancies, lymphoma-specific immune-related criteria have been developed (LYRIC), and they have been evaluated in research studies in comparison to the Lugano Classification. In this review work, we illustrate the evolution of the response criteria in lymphomas from the first CT-based criteria to the development of the PET-based Lugano Classification, further refined to take into account the flare phenomenon encountered during immunotherapy. We also describe the additional contribution of PET-derived volumetric parameters to the interpretation of responses during immunotherapy.

Methodological Aspects and the Prognostic Value of Metabolic Tumor Volume assessed with 18F-FDG PET/CT in Lymphomas.

Although metabolic tumor volume (MTV) assessed with pretreatment 18F-FDG PET/CT has shown significant prognostic value across many lymphoma types, it is still not used in clinical practice due to technical concerns and the lack of standardisation. Numerous studies on the prognostic value of MTV in lymphomas have been published in recent years, but there is still no full agreement on the best methodology for MTV calculation. In this paper, we reviewed the methodological aspects of MTV assessment and reported recent works about its impact on outcome in lymphomas, with a focus on Hodgkin lymphoma (HL) and diffuse large B cell lymphoma (DLBCL).

Semi-quantitative indices of cardiac uptake in patients with suspected cardiac amyloidosis undergoing 99mTc-HMDP scintigraphy.

BACKGROUND: 99mTc-HMDP scintigraphy has proved its efficacy in non-invasive diagnosis of cardiac amyloidosis (CA) and is currently interpreted according to the Perugini qualitative assessment. Several semi-quantitative indices have been proposed to overcome inherent possible limitations of visual grading. Our aim was to comparatively evaluate six different indices and their diagnostic performance. METHODS: We retrospectively reviewed scintigraphy of 76 patients (53 ATTR, 12 AL, 11 LVH) who underwent diagnostic evaluation at our centre. ROC-curve analysis was performed to identify optimal cut-off and relative diagnostic accuracy of six different indices (of which one was proposed for the first time), both in identifying CA patients and in discriminating patients according to their Perugini score. RESULTS: Heart/Whole-body ratios proved to be the most accurate (100%) in identifying CA patients. Heart/Pelvis ratio (with soft tissue background correction) offered acceptable accuracy (98%), with the largest area under the curve (AUC) (0.98) in discriminating patients with Perugini ≥ 2. Heart/Contralateral Lung ratio confirmed to be exposed to confounding background noise in case of simultaneous lung uptake. Heart/Skull ratio had the worst performance, with six false-negative patients in ATTR identification. CONCLUSION: Heart/Whole-body ratios may be robust and effective semi-quantitative indices for the evaluation of CA by means of scintigraphy.

Baseline metabolic tumor volume calculation using different SUV thresholding methods in Hodgkin lymphoma patients: interobserver agreement and reproducibility across software platforms.

AIM: Although it is not yet used in clinical practice, metabolic tumor volume (MTV) assessed on the baseline FDG-PET has shown consistent prognostic value in various lymphoma types. The aim of our study was to compare interobserver agreement and reproducibility across platforms of MTV calculation using different SUV thresholding methods in a large series of patients with newly diagnosed Hodgkin lymphoma. MATERIALS AND METHODS: We retrospectively studied 121 patients. MTV at baseline FDG-PET was independently computed by three readers with three programs of semi-automatic segmentation, Fiji, LifeX, and Accurate. MTV measurement was performed with different thresholds: SUV >2.5, SUV >4, and SUV >41% of SUV max. RESULTS: At inter-observer agreement analysis all Intraclass Correlation Coefficients (ICCs) were excellent (ICC >0.9), except for Accurate SUV >41% of SUV max (ICC = 0.8). The highest correlations were obtained at the SUV >4 threshold. The second best was SUV >2.5 threshold. Regarding reproducibility across software, we found statistically significant differences between Fiji versus LifeX and Accurate at fixed thresholds and between LifeX and Accurate at SUV >41% of SUV max, while no significant differences emerged between LifeX and Accurate using fixed thresholds. CONCLUSION: The three SUV thresholds studied are all suitable for MTV calculation in terms of reproducibility. The best reproducibility is achieved using fixed thresholds, both SUV >4 and SUV >2.5. If more than one software has to be used in a study, we suggest the use of fixed thresholds and the platforms LifeX and Accurate.

Accuracy of 99mTc-Hydroxymethylene diphosphonate scintigraphy for diagnosis of transthyretin cardiac amyloidosis.

BACKGROUND AND AIM: Either 99mTechnetium diphosphonate (Tc-DPD) or pyrophosphate (Tc-PYP) scintigraphy plays a relevant role in diagnosing transthyretin cardiac amyloidosis (CA), and labeled radiotracers have been extensively studied in diagnosing CA. Few studies have analyzed and validated 99mTc-Hydroxymethylene diphosphonate (Tc-HMDP). Our aim was to validate the diagnostic accuracy of Tc-HMDP total-body scintigraphy in a cohort of patients with biopsy-proven transthyretin CA. METHODS AND RESULTS: We retrospectively evaluated all patients undergoing 99mTc-HMDP total-body scintigraphy, in adjunct to a comprehensive diagnostic work-up for suspected CA. Sixty-five patients were finally diagnosed with CA, while it was excluded in 20 subjects with left ventricular hypertrophy of various etiologies. Twenty-six patients had AL-CA, 39 had TTR CA (16 TTRm, 23 TTRwt). At Tc-HMDP scintigraphy, 2 AL patients showed a Perugini score grade 1 heart uptake, while 24 showed no uptake. All TTR patients showed Tc-HMDP uptake, with three patients showing a Perugini score grade 1, 16 grade 2, and 20 grade 3, respectively. No uptake was observed in patients with left ventricular hypertrophy. A positive Tc-HMDP scintigraphy showed a 100% sensitivity and a 96% specificity for TTR CA identification. CONCLUSIONS: Tc-HMDP scintigraphy is as accurate as Tc-DPD or Tc-PYP, and may therefore de facto be considered a valuable tool for the diagnosis of TTR CA.

Lung uptake during 99mTc-hydroxymethylene diphosphonate scintigraphy in patient with TTR cardiac amyloidosis: An underestimated phenomenon.

BACKGROUND: Full body scintigraphy using bone tracers plays an important role in defining the type of amyloidosis and in diagnosing the heart involvement (cardiac amyloidosis, CA). No study has been conducted to explore lung retention (LR) in CA and its correlation to heart retention (HR).We evaluated LR in patients undergoing 99mTc-HMDP scintigraphy during evaluation for suspected CA. METHODS AND RESULTS: We enrolled 93 suspected CA patients. Patients underwent a complete diagnostic work up. After diagnostic process 82 patients resulted affected by certain CA (20 AL and 62 TTR), while 11 subjects showed left ventricular hypertrophy (LVH) not caused by CA. 99mTc-HMDP cardiac uptake was evaluated using the Perugini visual score while the modified Janssen score was used for LR estimation (grade 0 no uptake, grade 1 less than ribs, grade 2 more than ribs). RESULTS: 99mTc-HMDP LR was observed in 1/20 AL patient (5%), while 36/62 (58%) TTR patients showed LR with 29 grade 1 (47%) and 7 grade 2 (11%). No LR was observed in patients with LVH and no CA. LR was not evident in patients without HR, present in 1/3 (33%) of the patients with Perugini 1 HR and 11/24 (46%) and 26/36 (72%) of the patients showing respectively a Perugini 2 and a Perugini 3. CONCLUSION: 99mTc-HMDP scintigraphy shows LR in about 60% of TTR subjects, related to the grade of HR. In AL amyloidosis LR is less frequent than in TTR amyloidosis suggesting an aetiological tropism that seems comparable to the already known TTR related cardiac tropism.

Prediction of functional recovery after primary PCI using the estimate of myocardial salvage in gated SPECT early after acute myocardial infarction.

PURPOSE: Primary percutaneous coronary intervention (PCI) in acute myocardial infarction (AMI) aims to achieve myocardial salvage (MS). Because the reference method for measuring MS requires myocardial perfusion imaging (MPI) after tracer injection before PCI, alternative approaches have been proposed, but none has gained wide acceptance. Gated SPECT MPI can assess infarct size (IS), but can also show myocardial stunning. Thus, we compared functional and perfusion abnormalities early after AMI to estimate MS, and to predict left ventricular ejection fraction (LVEF) recovery at follow-up. METHODS: We studied 120 patients with AMI. Gated SPECT MPI was performed early (before hospital discharge) and at 6 months after AMI to measure IS, MS and functional outcome. MS was defined as the difference between the number of segments with abnormal thickening (i.e. the stunned area or area at risk) and the number of segments with abnormal perfusion (i.e. the final IS), expressed as a percentage of the total number of segments in the AHA model. LVEF was calculated using quantitative gated SPECT. RESULTS: The area at risk was 40 ± 25%, IS was 17.3 ± 16% and MS was 22 ± 19%. Early LVEF was 46.6 ± 11.6% and late LVEF was 51.4 ± 11.6%, with 54 patients showing at least an increase in LVEF of more than 5 units. ROC analysis showed that MS was able to predict LVEF recovery with an area under the curve (AUC) of 0.79 (p < 0.0001), and using a cut off >23% detected LVEF recovery with 74% sensitivity and 71% specificity. Conversely, IS was associated with an AUC 0.53 (not significant). CONCLUSION: MS assessed by a single early gated SPECT MPI study can accurately predict LVEF evolution after primary PCI for AMI.