Performance of automatic image segmentation algorithms for calculating total lesion glycolysis for early response monitoring in non-small cell lung cancer patients during concomitant chemoradiotherapy

Glycolytic Genes Predict Immune Status and Prognosis Non-Small-Cell Lung Cancer Patients with Radiotherapy and Chemotherapy

Background

Non-small-cell lung cancer (NSCLC) is a major health problem that endangers human health. The prognosis of radiotherapy or chemotherapy is still unsatisfactory. This study is aimed at investigating the predictive value of glycolysis-related genes (GRGs) on the prognosis of NSCLC patients with radiotherapy or chemotherapy.

Methods

Download the clinical information and RNA data of NSCLC patients receiving radiotherapy or chemotherapy from TCGA and geo databases and obtain GRGs from MsigDB. The two clusters were identified by consistent cluster analysis, the potential mechanism was explored by KEGG and GO enrichment analyses, and the immune status was evaluated by estimate, TIMER, and quanTIseq algorithms. Lasso algorithm is used to build the corresponding prognostic risk model.

Results

Two clusters with different GRG expression were identified. The high-expression subgroup had poor overall survival. The results of KEGG and GO enrichment analyses suggest that the differential genes of the two clusters are mainly reflected in metabolic and immune-related pathways. The risk model constructed with GRGs can effectively predict the prognosis. The nomogram combined with the model and clinical characteristics has good clinical application potential.

Conclusion

In this study, we found that GRGs are associated with tumor immune status and can assess the prognosis of NSCLC patients receiving radiotherapy or chemotherapy.

Development of Radiomic-Based Model to Predict Clinical Outcomes in Non-Small Cell Lung Cancer Patients Treated with Immunotherapy

Purpose: We aimed to assess the ability of radiomics features extracted from baseline (PET/CT0) and follow-up PET/CT scans, as well as their evolution (delta-radiomics), to predict clinical outcome (durable clinical benefit (DCB), progression, response to therapy, OS and PFS) in non-small cell lung cancer (NSCLC) patients treated with immunotherapy. Methods: 83 NSCLC patients treated with immunotherapy who underwent a baseline PET/CT were retrospectively included. Response was assessed at 6−8 weeks (PET/CT1) using PERCIST criteria and at 3 months with iPERCIST (PET/CT2) or RECIST 1.1 criteria using CT. The predictive performance of clinical parameters (CP), standard PET metrics (SUV, Metabolic Tumor volume, Total Lesion Glycolysis), delta-radiomics and PET and CT radiomics features extracted at baseline and during follow-up were studied. Seven multivariate models with different combinations of CP and radiomics were trained on a subset of patients (75%) using least absolute shrinkage, selection operator (LASSO) and random forest classification with 10-fold cross-validation to predict outcome. Model validation was performed on the remaining patients (25%). Overall and progression-free survival was also performed by Kaplan−Meier survival analysis. Results: Numerous radiomics and delta-radiomics parameters had a high individual predictive value of patient outcome with areas under receiver operating characteristics curves (AUCs) >0.80. Their performance was superior to that of CP and standard PET metrics. Several multivariate models were also promising, especially for the prediction of progression (AUCs of 1 and 0.96 for the training and testing subsets with the PET-CT model (PET/CT0)) or DCB (AUCs of 0.85 and 0.83 with the PET-CT-CP model (PET/CT0)). Conclusions: Delta-radiomics and radiomics features extracted from baseline and follow-up PET/CT images could predict outcome in NSCLC patients treated with immunotherapy and identify patients who would benefit from this new standard. These data reinforce the rationale for the use of advanced image analysis of PET/CT scans to further improve personalized treatment management in advanced NSCLC.

The dynamics and prognostic value of FDG PET-metrics in weekly monitoring of (chemo)radiotherapy for NSCLC

PURPOSE

To test if the relative change in FDG-PET SUV_max over the course of treatment was associated with disease progression and overall survival. Additionally, the prognostic values of other first-order PET-metric changes were investigated.

METHODS

The study included 38 patients with stage II-III NSCLC, who underwent concurrent chemoradiotherapy. Patients received two pre-treatment FDG-PET scans and four during-treatment scans at weekly intervals. SUV_max was normalized to the start of treatment and analyzed using linear regression. Linear regression coefficients of other first order PET-metrics were grouped according to dissimilarity. Associations to patient outcome were analyzed using Cox hazard ratio.

RESULTS

Twenty-eight patients satisfied the criteria for analysis. All PET-metrics demonstrated a strong linear correlation with time during treatment [median R-range: -0.87: -0.97]. No strong associations (p > 0.10) were found for the relative slope of SUV_max to patient outcomes. Other first-order metrics did correlate with outcome but the single imaging time-point maximizing the association of PET response with outcome varied per PET metric and outcome parameter.

CONCLUSION

All investigated FDG PET metrics linearly decreased during treatment. Relative change in SUV_max was not associated to patient outcome while several other first order PET-metrics were related to patient outcome. A single optimal imaging time-point could not be identified.

Convolutional neural networks for PET functional volume fully automatic segmentation: development and validation in a multi-center setting

Purpose

In this work, we addressed fully automatic determination of tumor functional uptake from positron emission tomography (PET) images without relying on other image modalities or additional prior constraints, in the context of multicenter images with heterogeneous characteristics.

Methods

In cervical cancer, an additional challenge is the location of the tumor uptake near or even stuck to the bladder. PET datasets of 232 patients from five institutions were exploited. To avoid unreliable manual delineations, the ground truth was generated with a semi-automated approach: a volume containing the tumor and excluding the bladder was first manually determined, then a well-validated, semi-automated approach relying on the Fuzzy locally Adaptive Bayesian (FLAB) algorithm was applied to generate the ground truth. Our model built on the U-Net architecture incorporates residual blocks with concurrent spatial squeeze and excitation modules, as well as learnable non-linear downsampling and upsampling blocks. Experiments relied on cross-validation (four institutions for training and validation, and the fifth for testing).

Results

The model achieved good Dice similarity coefficient (DSC) with little variability across institutions (0.80 ± 0.03), with higher recall (0.90 ± 0.05) than precision (0.75 ± 0.05) and improved results over the standard U-Net (DSC 0.77 ± 0.05, recall 0.87 ± 0.02, precision 0.74 ± 0.08). Both vastly outperformed a fixed threshold at 40% of SUVmax (DSC 0.33 ± 0.15, recall 0.52 ± 0.17, precision 0.30 ± 0.16). In all cases, the model could determine the tumor uptake without including the bladder. Neither shape priors nor anatomical information was required to achieve efficient training.

Conclusion

The proposed method could facilitate the deployment of a fully automated radiomics pipeline in such a challenging multicenter context.

Supplementary Information

The online version contains supplementary material available at 10.1007/s00259-021-05244-z.

A PET/CT volumetric parameter predicts prognosis of non-small cell lung cancer treated using preoperative chemoradiotherapy and surgery: A retrospective case series study

The purpose of the present study was to clarify whether positron emission tomography/computed tomography (PET/CT) volumetric parameters were prognostic predictors of non-small cell lung cancer (NSCLC) treatment in patients who had undergone preoperative concurrent chemoradiotherapy (CCRT) and surgery. In the present study, retrospectively surveyed the data of patients with NSCLC who underwent preoperative CCRT and surgery at Okayama University Hospital (Okayama, Japan) between April 2006 and March 2018. The maximum standardized uptake value (SUV_max) and volumetric parameters, including metabolic tumor volume (MTV) and total lesion glycolysis (TLG), were calculated using PET/CT and the percentage decrease (Δ) in each parameter value post-CCRT. The SUV_max threshold for defining MTV was set at 2.5. Furthermore, the association between survival and PET parameter values was analyzed. A total of 52 patients were included in the present study. The median follow-up period was 50.65 months. In univariate analysis, ΔTLG was identified to be a significant predictor of progression-free survival (PFS; P=0.03). The 5-year PFS rates were 48.6 and 76.6% for patients with low ΔTLG and high ΔTLG, respectively. High ΔTLG was indicative of a higher overall survival rate (P=0.08). The present results suggest that ΔTLG calculated using PET/CT is a prognostic predictor of NSCLC treated using preoperative CCRT and surgery, and may help physicians determine treatment strategies.

Local and regional treatment response by 18FDG-PET-CT-scans 4 weeks after concurrent hypofractionated chemoradiotherapy in locally advanced NSCLC

BACKGROUND AND PURPOSE

To investigate associations of early post-treatment ¹⁸Fluorodeoxyglucose-positron-emission-tomography (FDG-PET)-scans with local (LF), regional (RF), distant failure (DF) and overall survival (OS) in locally advanced non-small cell lung cancer (LA-NSCLC)-patients treated with concurrent chemoradiotherapy.

MATERIALS AND METHODS

Forty-seven stage IIIA-B NSCLC-patients included in a randomized phase II-trial (NTR2230) received 66 Gy (24x2.75 Gy) with low dose Cisplatin +/- Cetuximab. FDG-PET-scans were performed at baseline and 4 weeks post-treatment (range, 1.6-10.1). SUV_max, SUV_mean, metabolic tumor volume (MTV), total lesion glycolysis (TLG) and gross tumor volume were calculated separately for the primary tumor and the involved lymph nodes to generate baseline, post-treatment, and relative response metrics defined as (metric_pre-metric_post)/metric_pre. Univariable cox regression analyses were performed to investigate associations between PET-metrics and outcomes.

RESULTS

Metrics resulted from the post-treatment scan and relative response were associated with outcome, but baseline metrics were not. Primary tumor metrics were stronger associated with all outcomes than lymph node metrics. Both the volumetric (TLG/MTV) and intensity (SUV_max/SUV_mean) PET-metrics were associated with OS. The intensity metrics were associated with LF, while the volumetric PET-metrics were associated with RF/DF. This was in contrast to the nodal metrics, demonstrating only an association between RF and the relative response of TLG/MTV. No preference was found between PET volumetric and intensity metrics associated with outcome.

CONCLUSION

Early post-treatment PET-metrics are associated with treatment outcome in LA-NSCLC patients treated with chemoradiotherapy. Both volumetric and intensity PET-metrics are useful, but more for the primary tumor than for lymph nodes.

Improving the Spatial Alignment in PET/CT Using Amplitude-Based Respiration-Gated PET and Patient-Specific Breathing-Instructed CT

Appropriate attenuation correction is important for accurate quantification of SUVs in PET. Patient respiratory motion can introduce a spatial mismatch between respiration-gated PET and CT, reducing quantitative accuracy. In this study, the effect of a patient-specific breathing-instructed CT protocol on the spatial alignment between CT and amplitude-based optimal respiration-gated PET images was investigated. Methods: ¹⁸F-FDG PET/CT imaging was performed on 20 patients. In addition to the standard low-dose free-breathing CT, breath-hold CT was performed. The amplitude limits of the respiration-gated PET were used to instruct patients to hold their breath during CT acquisition at a similar amplitude level. Spatial mismatch was quantified using the position differences between the lung-liver transition in PET and CT images, the distance between PET and CT lesions' centroids, and the amount of overlap as indicated by the Jaccard similarity coefficient. Furthermore, the effect on attenuation correction was quantified by measuring SUVs, metabolic tumor volume, and total lesion glycolysis (TLG) of lung lesions. Results: All patients found the breathing instructions feasible; however, 4 patients had trouble complying with the instructions. In total, 18 patients were included. The average distance between the lung-liver transition between PET and CT was significantly reduced for breath-hold CT (1.7 ± 2.1 mm), compared with standard CT (5.6 ± 7.3 mm) (P = 0.049). Furthermore, the mean distance between the lesions' centroids on PET and CT was significantly smaller for breath-hold CT (3.6 ± 2.0 mm) than for standard CT (5.5 ± 6.5 mm) (P = 0.040). Quantification of lung lesion SUV was significantly affected, with a higher SUV_mean when breath-hold CT (6.3 ± 3.9 g/cm³) was used for image reconstruction than for standard CT (6.1 ± 3.8 g/cm³) (P = 0.044). Though metabolic tumor volume was not significantly different, TLG reached statistical significance. Conclusion: Optimal respiration-gated PET in combination with patient-specific breathing-instructed CT results in an improved alignment between PET and CT images and shows an increased SUV_mean and TLG. Even though the effects are small, a more accurate SUV and TLG determination is of importance for a more stable PET quantification, which is relevant for radiotherapy planning and therapy response monitoring.

Image-guided adaptive radiotherapy in patients with locally advanced non-small cell lung cancer: the art of PET

With a worldwide annual incidence of 1.8 million cases, lung cancer is the most diagnosed form of cancer in men and the third most diagnosed form of cancer in women. Histologically, 80-85% of all lung cancers can be categorized as non-small cell lung cancer (NSCLC). For patients with locally advanced NSCLC, standard of care is fractionated radiotherapy combined with chemotherapy. With the aim of improving clinical outcome of patients with locally advanced NSCLC, combined and intensified treatment approaches are increasingly being used. However, given the heterogeneity of this patient group with respect to tumor biology and subsequent treatment response, a personalized treatment approach is required to optimize therapeutic effect and minimize treatment induced toxicity. Medical imaging, in particular positron emission tomography (PET), before and during the course radiotherapy is increasingly being used to personalize radiotherapy. In this setting, PET imaging can be used to improve delineation of target volumes, employ molecularly-guided dose painting strategies, early response monitoring, prediction and monitoring of treatment-related toxicity. The concept of PET image-guided adaptive radiotherapy (IGART) is an interesting approach to personalize radiotherapy for patients with locally advanced NSCLC, which might ultimately contribute to improved clinical outcomes and reductions in frequency of treatment-related adverse events in this patient group. In this review, we provide a comprehensive overview of available clinical data supporting the use of PET imaging for IGART in patients with locally advanced NSCLC.

Role of interim 18F-FDG-PET/CT for the early prediction of clinical outcomes of Non-Small Cell Lung Cancer (NSCLC) during radiotherapy or chemo-radiotherapy. A systematic review

BACKGROUND

Non-Small Cell Lung Cancer (NSCLC) is characterized by aggressiveness and includes the majority of thorax malignancies. The possibility of early stratification of patients as responsive and non-responsive to radiotherapy with a non-invasive method is extremely appealing. The distribution of the Fluorodeoxyglucose (¹⁸F-FDG) in tumours, provided by Positron-Emission-Tomography (PET) images, has been proved to be useful to assess the initial staging of the disease, recurrence, and response to chemotherapy and chemo-radiotherapy (CRT).

OBJECTIVES

In the last years, particular efforts have been focused on the possibility of using ad interim ¹⁸F-FDG PET (FDG_int) to evaluate response already in the course of radiotherapy. However, controversial findings have been reported for various malignancies, although several results would support the use of FDG_int for individual therapeutic decisions, at least in some pathologies. The objective of the present review is to assemble comprehensively the literature concerning NSCLC, to evaluate where and whether FDG_int may offer predictive potential.

METHODS

Several searches were completed on Medline and the Embase database, combining different keywords. Original papers published in the English language from 2005 to 2016 with studies involving FDG_int in patients affected by NSCLC and treated with radiation therapy or chemo-radiotherapy only were chosen.

RESULTS

Twenty-one studies out of 970 in Pubmed and 1256 in Embase were selected, reporting on 627 patients.

CONCLUSION

Certainly, the lack of univocal PET parameters was identified as a major drawback, while standardization would be required for best practice. In any case, all these papers denoted FDG_int as promising and a challenging examination for early assessment of outcomes during CRT, sustaining its predictivity in lung cancer.

Practical no-gold-standard evaluation framework for quantitative imaging methods: application to lesion segmentation in positron emission tomography

Recently, a class of no-gold-standard (NGS) techniques have been proposed to evaluate quantitative imaging methods using patient data. These techniques provide figures of merit (FoMs) quantifying the precision of the estimated quantitative value without requiring repeated measurements and without requiring a gold standard. However, applying these techniques to patient data presents several practical difficulties including assessing the underlying assumptions, accounting for patient-sampling-related uncertainty, and assessing the reliability of the estimated FoMs. To address these issues, we propose statistical tests that provide confidence in the underlying assumptions and in the reliability of the estimated FoMs. Furthermore, the NGS technique is integrated within a bootstrap-based methodology to account for patient-sampling-related uncertainty. The developed NGS framework was applied to evaluate four methods for segmenting lesions from F-Fluoro-2-deoxyglucose positron emission tomography images of patients with head-and-neck cancer on the task of precisely measuring the metabolic tumor volume. The NGS technique consistently predicted the same segmentation method as the most precise method. The proposed framework provided confidence in these results, even when gold-standard data were not available. The bootstrap-based methodology indicated improved performance of the NGS technique with larger numbers of patient studies, as was expected, and yielded consistent results as long as data from more than 80 lesions were available for the analysis.