Radiomics, a growing area of cancer research that extracts non-invasive biomarkers from medical images, may improve lung cancer screening by identifying early-stage patients who are at high risk of poor prognosis.
It uses the National Lung Screening Trial (NLST) data to develop and validate a radiomics-based model that can identify vulnerable high-risk groups in early-stage patients associated with poor outcomes. The conclusion. These patients generally require active follow-up and/or adjuvant therapy.
Research Release June 29 Nature Scientific Reports..
Radiomics, also known as quantitative imagery, is a non-invasive biomarker generated from medical images. A new area of translational research, radiomics, extracts a large number of features from radiological images using data characterization algorithms that reflect the pathophysiology and heterogeneity of the underlying tumor.
These quantitative image features are rapidly calculated from standard-treatment images, reflect the overall tumor burden, and are not the only sample as in the case of tissue, so circulation and tissue-based bioassays for radiomics. The authors state that there are many advantages over markers based biomarkers.
“We see radiomics as a suite of decision-support tools for cancer management, whether screening and early detection, diagnosis, prognosis, or treatment response,” H. Cancer Epidemiology. Matthew B., the first author who is an associate member of. Remofite Cancer Center and Research Center in Tampa, Florida.
“Radiation features are generated from standard treatment images, and validated radiation models can provide clinicians with real-time decision support information,” he explained.
last year, Another study Showed that a combination of radiomics and imaging could identify patients with lung cancer who are most likely to respond to chemotherapy. The researchers used CT imaging of radiation features from inside and outside of lung nodules and found that it could predict time to patient progression and overall survival, as well as response to chemotherapy. Non-small cell lung cancer (NSCLC).
Dr. Anant Madabushi, professor of biomedical engineering and director of the Center for Computational Imaging and Personalized Diagnostic at Case Western Reserve University in Cleveland, Ohio, said the new study was “complementary, including radioactive material from both internal and external sources. It supports the premise of.” Tumors tell us about the outcome and treatment response. “
Madabushi also said that his group was on a similar line, Investigation Show how radiomics can predict the benefits of adjuvant therapy in lung cancer Investigation To show how radiomics can predict the recurrence of early NSCLC, Investigation We have shown that radiomics can predict NSCLC survival and response to immunotherapy.
Improving current lung cancer screening
Landmark NLST showed Compared with chest x-ray, low-dose helical computed tomography (LDCT) was associated with a 20% relative reduction in lung cancer mortality in high-risk individuals. However, LDCT screening Overdiagnosis and subsequent overtreatment Of slow-growing indolent cancer.
“The current selection criteria for lung cancer screening in the United States are primarily based on the criteria used by NLST,” Schabath said. Medscape Medical News.. “NLST has clearly shown that screening LDCTs is a life-saving tool, but NLST is not designed to create public policy.”
He found that less than 30% of Americans diagnosed with lung cancer met current screening criteria, and subsequent trials (eg, NELSON, LUSI, or MILD) used broader and more comprehensive criteria for early detection. He also pointed out that the efficacy of LDCT for lung cancer was also shown. “Therefore, there should be consideration in making lung cancer screening guidelines more comprehensive,” Schabath said.
“Additionally, supplemental risk stratification tools such as blood-based biomarkers can be an important complement to who should participate in lung cancer screening programs,” he said. “This is especially noticeable for people who have no or very few risk factors, such as those who have never smoked.”
Identifying bad results
In the current study, Schabath and colleagues used radiation data from NLST and LDCT images to generate radiation features from accidentally diagnosed lung cancer detected on screen. Next, radiomic features that account for size, shape, volume, and texture properties were calculated from both intratumoral and peritumoral regions.
Patients were divided into a training cohort and a test cohort, and an external cohort of non-screening-detected lung cancer patients was used for further validation. There were no statistically significant differences between most demographic training and test cohorts, including age, sex, smoking status, years of packs smoked, treatment, stage, and baseline screening results. However, self-report Chronic obstructive pulmonary disease (COPD) was significantly higher in the test cohort compared to the training group (16% vs. 7%, P = .02).
A total of 91 stable and reproducible radiomic functions (peritumoral and intratumoral) were identified and 40 (26 peritumoral and 14 intratumoral) were significantly associated with overall survival in the training cohort. The functionality was then narrowed down to four and backward exclusion analysis identified a single model. Patients were then divided into three risk groups: low risk, intermediate risk, and high risk.
High-risk groups had worse overall survival according to their model (hazard ratio [HR], 9.91; 25% 2.5 years, 0% 5 years OS), lower risk group (HR, 1.00, 93% 2.5 years, 78% 5 years OS).
The final model was validated in a test group and replicated in non-screen-detected patients with adenocarcinoma. Because the stages differed significantly across risk groups, the model was stratified by stage, and the authors found “convincing” results in early patients with generally good outcomes. In this subset, high-risk group (HR, 2.63, 56% 2.5 years and 42% 5 years OS) and low-risk group (HR, 1.00, 75% 2.5 years, 75% 5 years OS).
“We have ongoing research to determine if these results are consistent in a real-world setting for lung cancer screening across multiple centers,” Schabath said. “When NELSON, LUSI, or MILD trial data is published, we make sure that we seek to validate the results in these clinical trials.”
This study was funded by the National Cancer Institute. Schabath and Madabhushi do not disclose related financial relationships.
Person in charge.. Published June 29, 2020 Online. Full text