Association of radiomic features with genomic signatures in thyroid cancer: a systematic review.

Background: There is a growing interest on the association of radiomic features with genomic signatures in oncology. Using computational methods, quantitative radiomic data are extracted from various imaging techniques and integrated with genomic information to construct predictive models aimed at advancing diagnostic strategies in cancer patient management. In this context, the aim of this systematic review was to assess the current knowledge on potential application of this association in patients with thyroid cancer (TC).

Methods: A comprehensive literature review was conducted by querying three different databases (PubMed, Scopus and Embase) to identify studies published until June 2024, focusing on the potential association of radiomics and genomics in patients with TC. Pertinent data were subsequently extracted, and the methodological quality was evaluated using the A MeaSurement Tool to Assess Systematic Reviews 2 (AMSTAR 2).

Results: From the initial analysis, a total of 853 papers were identified. After removing duplicates and applying eligibility criteria, we ultimately evaluated 7 articles. It was observed that the most commonly utilized imaging technique for TC examination was ultrasound (US), followed by computed tomography and magnetic resonance imaging. Regarding genomic techniques, sequencing and polymerase chain reaction were the most commonly employed methods to validate genetic alterations. The association of radiomic features with genomic signatures demonstrated promising performance in predicting metastasis to the cervical lymph nodes or RET/PTC rearrangements. The effectiveness of models based on US-radiomic features in predicting BRAF mutation in patients with TC requires further investigation.

Conclusion: Although this systematic review has several limitations, primarily related to the limited amount of available literature data, the association of radiomic features with genomic signatures demonstrates a potential as non-invasive tool to enhance the accuracy and efficacy of TC diagnosis and prognosis. PROSPERO registration number: CRD42024572292.