Faculty Bibliography
Searched for:
person:jainr04
in-biosketch:true
Total Results:
212
MRI features predict tumor grade in isocitrate dehydrogenase (IDH)-mutant astrocytoma and oligodendroglioma
PURPOSE/OBJECTIVE:Nearly all literature for predicting tumor grade in astrocytoma and oligodendroglioma pre-dates the molecular classification system. We investigated the association between contrast enhancement, ADC, and rCBV with tumor grade separately for IDH-mutant astrocytomas and molecularly-defined oligodendrogliomas. METHODS:For this retrospective study, 44 patients with IDH-mutant astrocytomas (WHO grades II, III, or IV) and 39 patients with oligodendrogliomas (IDH-mutant and 1p/19q codeleted) (WHO grade II or III) were enrolled. Two readers independently assessed preoperative MRI for contrast enhancement, ADC, and rCBV. Inter-reader agreement was calculated, and statistical associations between MRI metrics and WHO grade were determined per reader. RESULTS:For IDH-mutant astrocytomas, both readers found a stepwise positive association between contrast enhancement and WHO grade (Reader A: OR 7.79 [1.97, 30.80], p = 0.003; Reader B: OR 6.62 [1.70, 25.82], p = 0.006); both readers found that ADC was negatively associated with WHO grade (Reader A: OR 0.74 [0.61, 0.90], p = 0.002); Reader B: OR 0.80 [0.66, 0.96], p = 0.017), and both readers found that rCBV was positively associated with WHO grade (Reader A: OR 2.33 [1.35, 4.00], p = 0.002; Reader B: OR 2.13 [1.30, 3.57], p = 0.003). For oligodendrogliomas, both readers found a positive association between contrast enhancement and WHO grade (Reader A: OR 15.33 [2.56, 91.95], p = 0.003; Reader B: OR 20.00 [2.19, 182.45], p = 0.008), but neither reader found an association between ADC or rCBV and WHO grade. CONCLUSIONS:Contrast enhancement predicts WHO grade for IDH-mutant astrocytomas and oligodendrogliomas. ADC and rCBV predict WHO grade for IDH-mutant astrocytomas, but not for oligodendrogliomas.
PMID: 35953567
ISSN: 1432-1920
CID: 5287192
Federated learning enables big data for rare cancer boundary detection
Although machine learning (ML) has shown promise across disciplines, out-of-sample generalizability is concerning. This is currently addressed by sharing multi-site data, but such centralization is challenging/infeasible to scale due to various limitations. Federated ML (FL) provides an alternative paradigm for accurate and generalizable ML, by only sharing numerical model updates. Here we present the largest FL study to-date, involving data from 71 sites across 6 continents, to generate an automatic tumor boundary detector for the rare disease of glioblastoma, reporting the largest such dataset in the literature (n = 6, 314). We demonstrate a 33% delineation improvement for the surgically targetable tumor, and 23% for the complete tumor extent, over a publicly trained model. We anticipate our study to: 1) enable more healthcare studies informed by large diverse data, ensuring meaningful results for rare diseases and underrepresented populations, 2) facilitate further analyses for glioblastoma by releasing our consensus model, and 3) demonstrate the FL effectiveness at such scale and task-complexity as a paradigm shift for multi-site collaborations, alleviating the need for data-sharing.
PMCID:9722782
PMID: 36470898
ISSN: 2041-1723
CID: 5381682
Periodic Alternating Gaze Deviation
PMID: 35421037
ISSN: 1536-5166
CID: 5204432
A Pipeline for the Implementation and Visualization of Explainable Machine Learning for Medical Imaging Using Radiomics Features
Machine learning (ML) models have been shown to predict the presence of clinical factors from medical imaging with remarkable accuracy. However, these complex models can be difficult to interpret and are often criticized as "black boxes". Prediction models that provide no insight into how their predictions are obtained are difficult to trust for making important clinical decisions, such as medical diagnoses or treatment. Explainable machine learning (XML) methods, such as Shapley values, have made it possible to explain the behavior of ML algorithms and to identify which predictors contribute most to a prediction. Incorporating XML methods into medical software tools has the potential to increase trust in ML-powered predictions and aid physicians in making medical decisions. Specifically, in the field of medical imaging analysis the most used methods for explaining deep learning-based model predictions are saliency maps that highlight important areas of an image. However, they do not provide a straightforward interpretation of which qualities of an image area are important. Here, we describe a novel pipeline for XML imaging that uses radiomics data and Shapley values as tools to explain outcome predictions from complex prediction models built with medical imaging with well-defined predictors. We present a visualization of XML imaging results in a clinician-focused dashboard that can be generalized to various settings. We demonstrate the use of this workflow for developing and explaining a prediction model using MRI data from glioma patients to predict a genetic mutation.
PMCID:9318445
PMID: 35890885
ISSN: 1424-8220
CID: 5276542
Federated Learning Enables Big Data for Rare Cancer Boundary Detection [PrePrint]
ORIGINAL:0015699
ISSN: 2331-8422
CID: 5284542
Neuroimaging of cerebrovascular complications in cancer patients
[S.l.] : Elsevier, 2022
pp. 935-954
ISBN: 9780128229958
CID: 5500792
Quantifying T2-FLAIR Mismatch Using Geographically Weighted Regression and Predicting Molecular Status in Lower-Grade Gliomas
BACKGROUND AND PURPOSE/OBJECTIVE:-mutant 1p/19q noncodeleted gliomas with a high positive predictive value. We have developed an approach to quantify the T2-FLAIR mismatch signature and use it to predict the molecular status of lower-grade gliomas. MATERIALS AND METHODS/METHODS:We used multiparametric MR imaging scans and segmentation labels of 108 preoperative lower-grade glioma tumors from The Cancer Imaging Archive. Clinical information and T2-FLAIR mismatch sign labels were obtained from supplementary material of relevant publications. We adopted an objective analytic approach to estimate this sign through a geographically weighted regression and used the residuals for each case to construct a probability density function (serving as a residual signature). These functions were then analyzed using an appropriate statistical framework. RESULTS:-mutant 1p/19q noncodeleted class of tumors versus other categories. Our classifier predicts these cases with area under the curve of 0.98 and high specificity and sensitivity. It also predicts the T2-FLAIR mismatch sign within these cases with an under the curve of 0.93. CONCLUSIONS:-mutation and 1p/19q codeletion status with high predictive power. The utility of the proposed quantification of the T2-FLAIR mismatch sign can be potentially validated through a prospective multi-institutional study.
PMID: 34764084
ISSN: 1936-959x
CID: 5050712
A Unified Approach to Analysis of MRI Radiomics of Glioma Using Minimum Spanning Trees
ISI:000797934200001
ISSN: 2296-424x
CID: 5284532
Prognostic indicators and outcomes of hospitalised COVID-19 patients with neurological disease: An individual patient data meta-analysis
BACKGROUND:Neurological COVID-19 disease has been reported widely, but published studies often lack information on neurological outcomes and prognostic risk factors. We aimed to describe the spectrum of neurological disease in hospitalised COVID-19 patients; characterise clinical outcomes; and investigate factors associated with a poor outcome. METHODS:We conducted an individual patient data (IPD) meta-analysis of hospitalised patients with neurological COVID-19 disease, using standard case definitions. We invited authors of studies from the first pandemic wave, plus clinicians in the Global COVID-Neuro Network with unpublished data, to contribute. We analysed features associated with poor outcome (moderate to severe disability or death, 3 to 6 on the modified Rankin Scale) using multivariable models. RESULTS:We included 83 studies (31 unpublished) providing IPD for 1979 patients with COVID-19 and acute new-onset neurological disease. Encephalopathy (978 [49%] patients) and cerebrovascular events (506 [26%]) were the most common diagnoses. Respiratory and systemic symptoms preceded neurological features in 93% of patients; one third developed neurological disease after hospital admission. A poor outcome was more common in patients with cerebrovascular events (76% [95% CI 67-82]), than encephalopathy (54% [42-65]). Intensive care use was high (38% [35-41]) overall, and also greater in the cerebrovascular patients. In the cerebrovascular, but not encephalopathic patients, risk factors for poor outcome included breathlessness on admission and elevated D-dimer. Overall, 30-day mortality was 30% [27-32]. The hazard of death was comparatively lower for patients in the WHO European region. INTERPRETATION:Neurological COVID-19 disease poses a considerable burden in terms of disease outcomes and use of hospital resources from prolonged intensive care and inpatient admission; preliminary data suggest these may differ according to WHO regions and country income levels. The different risk factors for encephalopathy and stroke suggest different disease mechanisms which may be amenable to intervention, especially in those who develop neurological symptoms after hospital admission.
PMCID:9162376
PMID: 35653330
ISSN: 1932-6203
CID: 5277632
Causal Inference in Radiomics: Framework, Mechanisms, and Algorithms
The widespread use of machine learning algorithms in radiomics has led to a proliferation of flexible prognostic models for clinical outcomes. However, a limitation of these techniques is their black-box nature, which prevents the ability for increased mechanistic phenomenological understanding. In this article, we develop an inferential framework for estimating causal effects with radiomics data. A new challenge is that the exposure of interest is latent so that new estimation procedures are needed. We leverage a multivariate version of partial least squares for causal effect estimation. The methodology is illustrated with applications to two radiomics datasets, one in osteosarcoma and one in glioblastoma.
PMCID:9261933
PMID: 35812228
ISSN: 1662-4548
CID: 5279682
