scholarly journals A machine learning‐based survival prediction model of high grade glioma by integration of clinical and dose‐volume histogram parameters

2021 ◽  
Vol 10 (8) ◽  
pp. 2774-2786
Author(s):  
Haiyan Chen ◽  
Chao Li ◽  
Lin Zheng ◽  
Wei Lu ◽  
Yanlin Li ◽  
...  
Keyword(s):  
Machine Learning ◽  
Prediction Model ◽  
High Grade Glioma ◽  
Survival Prediction ◽  
High Grade ◽  
Dose Volume Histogram ◽  
Dose Volume

RADT-20. SINGLE INSTITUTION REVIEW OF LARGE VOLUME HIGH GRADE GLIOMA PATIENTS TREATED WITH RADIOTHERAPY

2021 ◽  
Vol 23 (Supplement_6) ◽  
pp. vi45-vi45
Author(s):  
Vishal Manik ◽  
Angela Swampillai ◽  
Omar Al-Salihi ◽  
Kazumi Chia ◽  
Lucy Brazil
Keyword(s):  
Large Volume ◽  
Brain Volume ◽  
Collaborative Study ◽  
High Grade Glioma ◽  
Study Data ◽  
Molecular Profile ◽  
Normal Brain ◽  
High Grade ◽  
Total Brain Volume ◽  
Dose Volume

Abstract AIM Not uncommonly, we come across significantly large high grade glioma cases (HGGs). With standard delineation protocols, we end up irradiating a large volume of normal brain. Emami & QUANTEC data define normal brain tolerance doses, however they are often of limited use in clinic practice. Thus, we reviewed our patients with significant tumor volumes to derive a safe dose/ volume level for brain. METHODOLOGY Patients with HGGs over the last 3 years were extracted from Mosaiq™ information system. The output was sorted with respect to clinical target volumes from lowest to highest. The top 25 percentile i.e. patients with a CTV of > 412cc (n=53) were identified for this study. Data was collected with respect to clinical, tumor characteristics and radiotherapy parameters. RESULTS Median age of population was 53 and majority (n=38) were males. Nine patients had multi-focal tumors while six had bilateral extension. Majority of the study group had Glioblastoma Multiforme (n=44), whereas 6 had Grade 3 tumors. Most of the patients could only have a biopsy (n=27). Molecular profile showed 42 were Isocitrate-Dehydrogenase negative and 26 were unmethylated tumors. Stupp’s & Perry’s regimen were the commonly used protocols, however patients (n=7) with significant volumes near critical structures were treated with doses in the range of 50.4 – 55Gy in 30 fractions. The CTV volumes in the population ranged from 412 – 1223 cc while total brain volume range was 1112 – 1667 cc. Median of 43.5% of brain volume was covered in the PTV, while median of 5% of brain volume outside the PTV was treated to BED2 of 100Gy. Median survival was 12.4 months. CONCLUSION Our study shows reasonable tolerance of radiotherapy doses of > 50 Gy to larger volumes of brain. We propose a multi-center collaborative study to derive a new standardized dose volume tolerance.

scholarly journals RTHP-14. PSEUDOPROGRESSION IN HIGH GRADE GLIOMA PATIENTS AFTER PROTON OR PHOTON THERAPY

2019 ◽  
Vol 21 (Supplement_6) ◽  
pp. vi213-vi213
Author(s):  
Vonetta Williams ◽  
Lia Halasz ◽  
Jason Rockhill ◽  
James Fink
Keyword(s):  
Mr Imaging ◽  
High Grade Glioma ◽  
Extent Of Resection ◽  
High Dose ◽  
High Grade ◽  
Proton Group ◽  
Kaplan Meier ◽  
Dose Volume ◽  
Biological Effectiveness ◽  
Grade Iii

Abstract Pseudoprogression is defined as the appearance of false progression on MR imaging following radiation therapy. Proton therapy is thought to have increased relative biological effectiveness-the ratio of the doses required by two types of radiation to cause the same level of effect-near the edges of the high dose volume. This could lead to different rates of pseudoprogression for protons compared to photons. In our IRB approved study, a board-certified neuroradiologist reviewed serial imaging of 74 patients (photons: n=37, protons: n=37) treated from 2013–2018 with either proton or photon radiotherapy to 59.4–60 Gy in 30–33 fractions and temozolomide for high grade glioma. MR imaging was performed 1 month after completion of treatment and then every 3 months. True progression was scored based on updated RANO criteria. Pseudoprogression was determined if imaging improved without change in therapy. Cumulative incidences of these outcomes and survival were calculated utilizing Kaplan-Meier analyses. Patient and treatment factors were analyzed for their association with incidence of pseudoprogression. Median follow-up for alive patients in the proton and photon groups were 15 and 29 months, respectively. Median age was 49 years in the proton group and 54 years in the photon group (p=0.17). Among proton patients, 14 had grade III glioma and 23 had grade IV glioblastoma. Among photon patients, 1 had grade III glioma. Median survival was 23 and 35 months for the proton and photon groups, respectively (p=0.57). The cumulative incidence of pseudoprogression was 14.4% and 10.4% at 12 months for the proton and photon groups, respectively (p=0.53). Grade, extent of resection, age, and IDH status, were not significantly associated with development of pseudoprogression. MGMT methylated tumors showed a trend toward association with pseudoprogression compared to unmethylated tumors (p=0.058). We concluded that the incidence of pseudoprogression is similar regardless of whether proton or photon therapy was utilized.

scholarly journals Survival prediction based on qualitative MRI diffusion signature in patients with recurrent high grade glioma treated with bevacizumab

2018 ◽  
Vol 8 (3) ◽  
pp. 268-279 ◽  
Author(s):  
Pradeep Goyal ◽  
Mary Tenenbaum ◽  
Sonali Gupta ◽  
Puneet S. Kochar ◽  
Alok A. Bhatt ◽  
...  
Keyword(s):  
High Grade Glioma ◽  
Survival Prediction ◽  
High Grade

Combined structural and perfusion MRI enhanced by machine learning may outperform standalone modalities and radiological expertise in high-grade glioma surveillance: A proof-of-concept study.

2020 ◽  
Vol 38 (15_suppl) ◽  
pp. e14528-e14528
Author(s):  
Diana Roettger ◽  
Loizos Siakallis ◽  
Carole Sudre ◽  
Jasmina Panovska-Griffiths ◽  
Paul Mulholland ◽  
...  
Keyword(s):  
Machine Learning ◽  
High Grade Glioma ◽  
Treatment Monitoring ◽  
Perfusion Mri ◽  
Svm Classifier ◽  
High Grade ◽  
Time Points ◽  
Group I ◽  
Sensitivity Specificity ◽  
Time Point

e14528 Background: Treatment monitoring in patients with High-Grade Glioma (HGG) and identification of disease progression, remains a major challenge in clinical neurooncology. We aimed to develop a support vector machine (SVM) classifier utilising combined longitudinal conventional and Dynamic Susceptibility Contrast (DSC) perfusion MRI to classify between Stable Disease (SD), Pseudoprogression (PsP) and Progressive Disease (PD) in glioma patients under surveillance. Methods: Conventional (269) and perfusion (62) MRI studies of HGG patients acquired between 2012 and 2018 were prospectively analysed. Study participants were separated into two groups: Group I with a single DSC time point (64 participants) and Group II with multiple DSC time points (19 participants). The SVM classifier was trained using all available MRI for each group. Classification accuracy was assessed for the use of features extracted from different feature dataset and time point combinations and compared to the experienced radiologists’ predictions. Results: The study included 64 participants (mean age: 48.5 ± 12.8 yrs [standard deviation], 24 female). SVM classification based on combined perfusion and structural features outperformed standalone datasets across all groups. For the clinically relevant classification step (SD/PSP vs PD), both feature combination as well as the addition of multiple DSC time points, improved classification performance (lowest median error rate: 0.016). The SVM algorithm outperformed radiologists in predicting lesion destiny in both groups. Optimal performance was observed in Group II, in which SVM sensitivity/specificity/accuracy was 100/91.67/94.7% for analysis based on the first time point and 85.71/100/ 94.7% based on multiple time points compared to 60/78/68% and 70/90/84.2% for the respective radiologist classifications. In Group I, the SVM also exceeded radiologist predictions, albeit by a smaller margin and resulted in sensitivity/specificity of 86.49/75.00/81.53% (SVM) compared to 75.7/68.9/73.84% (radiologists). Conclusions: Our results indicate that the addition of multiple longitudinal perfusion time points as well as the combination of structural and perfusion features significantly enhance classification outcome in treatment monitoring of HGGs and machine-learning-assisted diagnosis has potentially superior accuracy than the radiologist's visual evaluation and expertise.

Sign in / Sign up

Export Citation Format

Share Document