# RadIO: A Novel Radiomics Toolkit to Predict and Characterize Response to Immunotherapy in Stage III Non-Small Cell Lung Cancer

> **NIH VA I01** · VETERANS HEALTH ADMINISTRATION · 2024 · —

## Abstract

PROJECT SUMMARY
In 2022, an estimated 236,740 patients in the US are expected to be diagnosed with non-small cell lung cancer
(NSCLC) and 130,180 deaths. In 2019 and 2020, the Veteran Health Administration estimates 8,352 and 6,111
new lung cancer diagnoses and 4,284 and 2,344 deaths from lung cancer, respectively. Management of Stage
III unresectable NSCLC underwent a monumental change with the NEJM publication of the PACIFIC trial in
2017. Maintenance immunotherapy after chemoradiation became the standard with median overall survival
(mOS) nearly doubling to 47.5 months compared to chemo-radiation alone. While with durvalumab, the
progression-free survival (PFS) rate at 60 months was 33.1%, compared to placebo (i.e. chemoradiation alone)
which was 19.0%, the findings suggest that 1 in 2 patients is likely to not receive added benefit from
durvalumab. The problem then becomes, who should receive immunotherapy over and above chemoradiation?
By identifying only those Veterans who will receive benefit will result in a significant cost savings for the VA
health care system. Second, many Veterans within the VA Healthcare System travel > 50 miles for their
treatments. By eliminating unnecessary trips for inefficacious therapy, the quality of life of our Veterans will be
enhanced. Similarly, given the cost and associated toxicity of these IO based regimens there is a need to identify
which patients will benefit from a short treatment duration (i.e. 6 months versus the typical 1 year duration).
Finally, there is a need to identify which patients are likely to suffer from immune related adverse events (irAEs),
specifically pneumonitis, seen in approximately 1/3 of lung cancer patients treated with IO.
 Our group has pioneered the development of novel artificial intelligence (AI) based radiomic (computer
extracted) features of the tortuosity of the nodule vasculature and textural patterns both inside and outside the
nodule on CT scans, and demonstrated strong association between these features with response to (1)
durvalumab+chemoradiation in Stage III NSCLC, (2) in first line chemo+IO and (3) overall survival in NSCLC
patients treated with IO. In recent work published in the J of Immunotherapy for Cancer (JITC), our group showed
that intra- and peritumoral radiomic texture features were associated with disease free survival (DFS) in Stage
III NSCLC patients treated with durvalumab+chemoradiation. The radiomic features were associated with clinical
outcome in both the low and high PD-L1 groups. More recently, in work to be presented at ASCO 2022, we have
shown that radiomic features on CT scans can distinguish radiation induced from IO induced pneumonitis.
 In this project, we aim to develop and validate RadIO, a novel radiomic toolkit to (1) predict at baseline
which Veterans will receive added benefit from durvalumab over chemo-radiation alone, (2) determine when a
disease stable status has been achieved so that an IO duration prediction can be made (6 ...

## Key facts

- **NIH application ID:** 10588483
- **Project number:** 2I01BX004121-06A1
- **Recipient organization:** VETERANS HEALTH ADMINISTRATION
- **Principal Investigator:** Anant Madabhushi
- **Activity code:** I01 (R01, R21, SBIR, etc.)
- **Funding institute:** VA
- **Fiscal year:** 2024
- **Award amount:** —
- **Award type:** 2
- **Project period:** 2019-01-01 → 2027-09-30

## Primary source

NIH RePORTER: https://reporter.nih.gov/project-details/10588483

## Citation

> US National Institutes of Health, RePORTER application 10588483, RadIO: A Novel Radiomics Toolkit to Predict and Characterize Response to Immunotherapy in Stage III Non-Small Cell Lung Cancer (2I01BX004121-06A1). Retrieved via AI Analytics 2026-05-25 from https://api.ai-analytics.org/grant/nih/10588483. Licensed CC0.

---

*[NIH grants dataset](/datasets/nih-grants) · CC0 1.0*