Population-Scale CT-based Body Composition Analysis of a Large                    Outpatient Population Using Deep Learning to Derive Age-, Sex-, and                    Race-specific Reference Curves

Magudia, Kirti; Bridge, Christopher P.; Bay, Camden; Babić, Ana; Fintelmann, Florian J.; Troschel, Fabian M.; Miskin, Nityanand; Wrobel, William C; Brais, Lauren K.; Andriole, Katherine P.; Wolpin, Brian M.; Rosenthal, Michael H.

doi:10.1148/radiol.2020201640

Cited by 88 publications

(66 citation statements)

References 40 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Our methods seek to address model interpretability, an important barrier to implementing artificial intelligence in medicine 51 . Though our segmentation model had high Dice scores, similar to other published studies 52,53 , BC metrics alone or in combination with PCE covariates did not outperform the PCE. To our knowledge, our approach is the first to model IHD risk prediction in an end-to-end manner using the pixel data, as opposed to BC metrics, outperforming the radiomics/PCE approach in 5-year IHD prediction.…”

Section: Discussionsupporting

confidence: 84%

Opportunistic Assessment of Ischemic Heart Disease Risk Using Abdominopelvic Computed Tomography and Medical Record Data: a Multimodal Explainable Artificial Intelligence Approach

et al. 2021

Preprint

View full text Add to dashboard Cite

Current risk scores for predicting ischemic heart disease (IHD) risk—the leading cause of global mortality—have limited efficacy. While body composition (BC) imaging biomarkers derived from abdominopelvic computed tomography (CT) correlate with IHD risk, they are impractical to measure manually. Here, in a retrospective cohort of 8,197 contrast-enhanced abdominopelvic CT examinations undergoing up to 5 years of follow-up, we developed improved multimodal opportunistic risk assessment models for IHD by automatically extracting BC features from abdominal CT images and integrating these with features from each patient’s electronic medical record (EMR). Our predictive methods match and, in some cases, outperform clinical risk scores currently used in IHD risk assessment. We provide clinical interpretability of our model using a new method of determining tissue-level contributions from CT along with weightings of EMR features contributing to IHD risk. We conclude that such a multimodal approach, which automatically integrates BC biomarkers and EMR data can enhance IHD risk assessment and aid primary prevention efforts for IHD.

show abstract

Section: Discussionsupporting

confidence: 84%

Opportunistic Assessment of Ischemic Heart Disease Risk Using Abdominopelvic Computed Tomography and Medical Record Data: a Multimodal Explainable Artificial Intelligence Approach

et al. 2021

Preprint

View full text Add to dashboard Cite

show abstract

“…Similar to prior studies, pulmonary function tests showed no significant effect 13 . As previously demonstrated, age was not associated with survival, highlighting that mortality risk is not so much defined by chronologic, but morphologic age 16,34–36 …”

Section: Discussionsupporting

confidence: 83%

Sarcopenia on preoperative chest computed tomography predicts cancer‐specific and all‐cause mortality following pneumonectomy for lung cancer: A multicenter analysis

et al. 2021

Self Cite

View full text Add to dashboard Cite

Background Mortality risk prediction in patients undergoing pneumonectomy for non‐small cell lung cancer (NSCLC) remains imperfect. Here, we aimed to assess whether sarcopenia on routine chest computed tomography (CT) independently predicts worse cancer‐specific (CSS) and overall survival (OS) following pneumonectomy for NSCLC. Methods We included consecutive adults undergoing standard or carinal pneumonectomy for NSCLC at Massachusetts General Hospital and Heidelberg University from 2010 to 2018. We measured muscle cross‐sectional area (CSA) on CT at thoracic vertebral levels T8, T10, and T12 within 90 days prior to surgery. Sarcopenia was defined as T10 muscle CSA less than two standard deviations below the mean in healthy controls. We adjusted time‐to‐event analyses for age, body mass index, Charlson Comorbidity Index, forced expiratory volume in 1 second in % predicted, induction therapy, sex, smoking status, tumor stage, side of pneumonectomy, and institution. Results Three hundred and sixty‐seven patients (67.4% male, median age 62 years, 16.9% early‐stage) underwent predominantly standard pneumonectomy (89.6%) for stage IIIA NSCLC (45.5%) and squamous cell histology (58%). Sarcopenia was present in 104 of 367 patients (28.3%). Ninety‐day all‐cause mortality was 7.1% (26/367). After a median follow‐up of 20.5 months (IQR, 9.2–46.9), 183 of 367 patients (49.9%) had died. One hundred and thirty‐three (72.7%) of these deaths were due to lung cancer. Sarcopenia was associated with shorter CSS (HR 1.7, p = 0.008) and OS (HR 1.7, p = 0.003). Conclusions This transatlantic multicenter study confirms that sarcopenia on preoperative chest CT is an independent risk factor for CSS and OS following pneumonectomy for NSCLC.

show abstract

“…Most of these studies focus on the segmentation of the tissue compartments in a single slice at a certain lumbar level, as it has been demonstrated that 2D and 3D measurements for quantification of VAT, SAT, and SM show a high correlation [9][10][11][12][13][14]. Although very recent works have also addressed automation of slice extraction, routine clinical application additionally requires the integration of quality control methods for both slice extraction and tissue segmentation [15,16]. For this purpose, two classic machine learning models have been developed in this study.…”

Section: Discussionmentioning

confidence: 99%

“…However, manual interaction was still required for extraction of the single slice on which the automatic segmentation is performed. Only very recent work also includes deeplearning-based automated slice extraction as the next step for truly automated body composition analyses [15][16][17].…”

Section: Introductionmentioning

confidence: 99%

End-to-end automated body composition analyses with integrated quality control for opportunistic assessment of sarcopenia in CT

et al. 2021

View full text Add to dashboard Cite

Objectives To develop a pipeline for automated body composition analysis and skeletal muscle assessment with integrated quality control for large-scale application in opportunistic imaging. Methods First, a convolutional neural network for extraction of a single slice at the L3/L4 lumbar level was developed on CT scans of 240 patients applying the nnU-Net framework. Second, a 2D competitive dense fully convolutional U-Net for segmentation of visceral and subcutaneous adipose tissue (VAT, SAT), skeletal muscle (SM), and subsequent determination of fatty muscle fraction (FMF) was developed on single CT slices of 1143 patients. For both steps, automated quality control was integrated by a logistic regression model classifying the presence of L3/L4 and a linear regression model predicting the segmentation quality in terms of Dice score. To evaluate the performance of the entire pipeline end-to-end, body composition metrics, and FMF were compared to manual analyses including 364 patients from two centers. Results Excellent results were observed for slice extraction (z-deviation = 2.46 ± 6.20 mm) and segmentation (Dice score for SM = 0.95 ± 0.04, VAT = 0.98 ± 0.02, SAT = 0.97 ± 0.04) on the dual-center test set excluding cases with artifacts due to metallic implants. No data were excluded for end-to-end performance analyses. With a restrictive setting of the integrated segmentation quality control, 39 of 364 patients were excluded containing 8 cases with metallic implants. This setting ensured a high agreement between manual and fully automated analyses with mean relative area deviations of ΔSM = 3.3 ± 4.1%, ΔVAT = 3.0 ± 4.7%, ΔSAT = 2.7 ± 4.3%, and ΔFMF = 4.3 ± 4.4%. Conclusions This study presents an end-to-end automated deep learning pipeline for large-scale opportunistic assessment of body composition metrics and sarcopenia biomarkers in clinical routine. Key Points • Body composition metrics and skeletal muscle quality can be opportunistically determined from routine abdominal CT scans. • A pipeline consisting of two convolutional neural networks allows an end-to-end automated analysis. • Machine-learning-based quality control ensures high agreement between manual and automatic analysis.

show abstract

Population-Scale CT-based Body Composition Analysis of a Large Outpatient Population Using Deep Learning to Derive Age-, Sex-, and Race-specific Reference Curves

Cited by 88 publications

References 40 publications

Opportunistic Assessment of Ischemic Heart Disease Risk Using Abdominopelvic Computed Tomography and Medical Record Data: a Multimodal Explainable Artificial Intelligence Approach

Opportunistic Assessment of Ischemic Heart Disease Risk Using Abdominopelvic Computed Tomography and Medical Record Data: a Multimodal Explainable Artificial Intelligence Approach

Sarcopenia on preoperative chest computed tomography predicts cancer‐specific and all‐cause mortality following pneumonectomy for lung cancer: A multicenter analysis

End-to-end automated body composition analyses with integrated quality control for opportunistic assessment of sarcopenia in CT

Contact Info

Product

Resources

About