Machine learning model can help reduce follow up screenings to 28% of people with benign nodules

A new machine learning model, developed in the Benos lab, can help reduce the unnecessary follow up screenings to people with benign nodules detected in low-dose CT scans by 28%.

Introduction Low-dose CT (LDCT) is currently used in lung cancer screening of high-risk populations for early lung cancer diagnosis. However, 96% of individuals with detected nodules are false positives.

Methods In order to develop an efficient early lung cancer predictor from clinical, demographic and LDCT features, we studied a total of 218 subjects with lung cancer or benign nodules. Probabilistic graphical models (PGMs) were used to integrate demographics, clinical data and LDCT features from 92 subjects (training cohort) from the Pittsburgh Lung Screening Study cohort.

Results Learnt PGMs identified three variables directly (causally) linked to malignant nodules and the largest benign nodule and used them to build the Lung Cancer Causal Model (LCCM), which was validated in a separate cohort of 126 subjects. Nodule and vessel numbers and years since the subject quit smoking were sufficient to discriminate malignant from benign nodules. Comparison with existing predictors in the training and validation cohorts showed that (1) incorporating LDCT scan features greatly enhances predictive accuracy; and (2) LCCM improves cancer detection over existing methods, including the Brock parsimonious model (p<0.001). Notably, the number of surrounding vessels, a feature not previously used in predictive models, significantly improves predictive efficiency. Based on the validation cohort results, LCCM is able to identify 30% of the benign nodules without risk of misclassifying cancer nodules.

Discussion LCCM shows promise as a lung cancer predictor as it is significantly improved over existing models. Validated in a larger, prospective study, it may help reduce unnecessary follow-up visits and procedures.

This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial.

Press:

UPMC: https://www.upmc.com/media/news/031219-lung-ca-machine-learning

NPR: https://www.wesa.fm/post/artificial-intelligence-could-reduce-false-positives-lung-cancer-screenings

Vineet K. Raghu

 

 

 

 

Takis Benos   

 

 

 

 

      David Wilson


Raghu, …, Benos*, Wilson. Feasibility of lung cancer prediction from low-dose CT scan and smoking factors using causal modelsThorax, 2019

Dr. Jingyu Zhang from Xing’s lab published a first-author paper in PLoS Computational Biology

Spatial clustering and common regulatory elements correlate with coordinated gene expression

Cellular responses to environmental stimulation are often accompanied by changes in gene expression patterns. Genes are linearly arranged along chromosomal DNA, which folds into a three-dimensional structure. The chromosome structure affects gene expression activities and is regulated by multiple events such as histone modifications and DNA binding of transcription factors. A basic question is how these mechanisms work together to regulate gene expression. In this study, we analyzed temporal gene expression patterns in the context of chromosome structure both in a human cell line under TGF-β treatment and during mouse nervous system development. In both cases, we observed that genes regulated by common transcription factors have an enhanced tendency to be spatially close. Our analysis suggests that spatial co-localization of genes may facilitate the concerted gene expression.

Zhang J, Chen H, Li R, Taft DA, Yao G, Bai F, Xing J. (2019) Spatial clustering and common regulatory elements correlate with coordinated gene expression. PLoS Computational Biology 15(3):e1006786

Congratulations to Drs. Nick Pabon and Qiuhong Zhang for their first author paper in Nature Comm.

Congratulations to Drs. Nick Pabon and Qiuhong Zhang for their first author paper in Nature Communications titled: A network-centric approach to drugging TNF-induced NF-kB signaling
with Profs. Carlos J. Camacho and Robin E.C. Lee.

Target-centric drug development strategies prioritize single-target potency in vitro and do not account for connectivity and multi-target effects within a signal transduction network. Here, we present a systems biology approach that combines transcriptomic and structural analyses with live-cell imaging to predict small molecule inhibitors of TNF-induced NF-κB signaling and elucidate the network response. We identify two first-in-class small molecules that inhibit the NF-κB signaling pathway by preventing the maturation of a rate-limiting multiprotein complex necessary for IKK activation. Our findings suggest that a network-centric drug discovery approach is a promising strategy to evaluate the impact of pharmacologic intervention in signaling.

 

Pabon NA, Zhang Q, Cruz JA, Schipper DL, Camacho CJ, Lee REC, (2019). A network-centric approach to drugging TNF-induced NF-κB signaling. Nature Communications. Vol 10, Article number: 860

Congratulations to Nick Pabon, PhD, for his first author paper in PlosCompBio

Dr. Nick Pabon publishes first author paper in PlosCompBio titled: “Predicting protein targets for drug-like compounds using transcriptomics” with Profs. Carlos J. Camacho and Ziv Bar-Joseph.

The paper describes the first method capable of predicting novel drug-target interactions using gene expression profiles. Nick contribution was specially notable in highlighting correlations between profiles indirectly related to the target as the main determinant of physical interactions in the machine learning model.

Read the full paper here