CT-based AI could be game changer for radiologists assessing invasive, noninvasive cancers
The utilization of a deep learning network could help radiologists differentiate between minimally invasive adenocarcinoma (MIA) and invasive adenocarcinoma (IAC) according to new research published this week in the European Journal of Radiology.
Distinguishing MIA from IAC based on CT images of subsolid pulmonary nodules (SSPNs) has historically been difficult, but the difference between the two diagnoses is critical for patients. The 5-year survival rate for patients with MIA is close to 100%, while those who have been diagnosed with IAC see that rate drop down to 40-85%.
“It is of critical importance to develop a prediction model to efficiently discriminate MIAs from IACs, thereby guiding treatment and predicting the prognosis of patients with SSPNs,” wrote author, Xianmeng Chen, with the Department of Radiology at Jiangmen Central Hospital, and co-authors.
The study included 365 patients treated at two medical centers from 2016-2019. All patients had SSPNs and pathologically confirmed MIA or IAC. Preoperative CT images were used to select deep learning features. The deep learning signature (DLS) was developed via the least absolute shrinkage and selection operator (LASSO).
Between the MIA and IAC groups, 18 learning features with non-zero coefficients were enrolled in the signature. Independent predictors of the DLS were used to help develop the deep learning network (DLN). In training, internal validation and external validation, the tool had AUCs of 0.89 0.94 and 0.91, respectively. Further analysis proved that the DLN was able to consistently discern between MIA and IAC.
The authors believe their research will help more accurately diagnose patients and direct appropriate treatments.
“In this study, we combined the DLS and subjective CT parameters to construct the DLN, which was a non-invasive, quantitative, and reproducible model that could be used to differentiate MIA from IAC sensitively and specifically,” the authors explained.
You can read the full study in the European Journal of Radiology.
In addition to her background in journalism, Hannah also has patient-facing experience in clinical settings, having spent more than 12 years working as a registered rad tech. She joined Innovate Healthcare in 2021 and has since put her unique expertise to use in her editorial role with Health Imaging.