Prognostic Factors in Adrenocortical Carcinoma: The Added Value of CT-Based Imaging Biomarkers
Canadian Association of Radiologists Journal 2025, Vol. 76(4) 568-569 @ The Author(s) 2025 Article reuse guidelines: sagepub.com/journals-permissions DOI: 10.1177/08465371251350605 journals.sagepub.com/home/caj
S Sage
Felipe Lopez-Ramirez! (D, Linda C. Chu’, and Elliot K. Fishman’ iD
Adrenocortical carcinoma (ACC) is an aggressive malig- nancy, marked by unpredictable biological behaviour, high recurrence rates, and limited response to systemic thera- pies. Surgical resection has traditionally been the corner- stone of ACC treatment, aiming to achieve a complete macroscopic resection to improve oncologic outcomes. However, even after complete surgical resection, disease recurrence is common, and predicting individual patient outcomes remains a significant clinical challenge. In this context, the ability to preoperatively assess the prognosis of ACC patients holds significant value for clinical practice, offering guidance to select treatment strategies and opti- mize therapeutic decision-making.
Computed tomography (CT) is the primary imaging modality for the evaluation of adrenal masses. It plays a critical role in differentiating ACC from other adrenal tumours and provides the foundation for preliminary clini- cal staging and guiding further functional and hormonal assessments. Building on its diagnostic utility, emerging studies have increasingly demonstrated the power of imag- ing features to inform clinical decisions such as differentiat- ing ACC tumours from benign entities.1 Quantitative features such as tumour shape, margins, texture, and enhancement patterns are now being explored as imaging biomarkers that may correlate with aggressive tumour char- acteristics and survival outcomes.2 As a result, CT is evolv- ing from a purely diagnostic tool into a powerful instrument for preoperative risk stratification, expanding its role beyond the initial morphological assessment.
Defining a “good prognosis” in ACC is inherently chal- lenging due to the biological heterogeneity of the disease and often unpredictable clinical course. Staging systems such as the ENSAT (European Network for the Study of Adrenal Tumours) and the AJCC (American Joint Committee on Cancer) provide valuable frameworks for categorizing dis- ease extent and guiding postoperative treatment decisions. A favorable prognosis has been linked to surgical and histopath- ological variables such as early-stage disease (ie, localized disease), complete surgical resection with negative margins, Ki-67 index, Weiss score, and the absence of a hormonal secretion syndrome.3 Since most staging factors largely rely on anatomical and pathologic features determined postopera- tively, patient prognostication in the preoperative setting
remains challenging and may fall short in accurately captur- ing individual risk to guide management.
In this issue of the Canadian Association of Radiologists Journal, Barat et al present a compelling analysis highlighting the prognostic value of the preoperative evaluation of ACC patients using CT features.4 In their timely proof-of-concept study, the authors explore the added value of shape radiomics features, such as tumour shape elongation and tumour maxi- mum diameter, integrated with established staging systems such as the ENSAT. These findings demonstrated that incor- porating CT-derived features significantly enhanced prognos- tic accuracy in surgically treated ACC patients. In an external validation of their model, the authors demonstrated an improvement in the concordance index (C-index) for overall survival prediction from a baseline 0.35 with the ENSAT staging alone, to a more robust 0.62 (P =. 002) with the CT-score model. The refined prognostic score also resulted in a clinically meaningful risk stratification, showing significant differences in overall survival times across the resulting risk groups. This work underscores the potential of imaging bio- markers to complement traditional staging and move the field toward more individualized, image-informed risk stratifica- tion in ACC.
This study aligns well with prior work using shape-based features for preoperative characterization of tumour aggres- siveness in ACC patients and shares common limitations, par- ticularly the small sample size reflecting the rarity of ACC tumours.1,2 Collaborative studies across institutions may be necessary to evaluate the performance of these features in diverse clinical and imaging scenarios, increase the statistical power of findings, and further elucidate the role of imaging biomarkers in ACC prognostication. Nonetheless, as dis- cussed by the authors, the labor-intensive manual segmenta- tion task required for radiomics feature extraction remains a significant barrier to the clinical implementation of these
“The Russell H. Morgan Department of Radiology and Radiological Science, Johns Hopkins University School of Medicine, Baltimore, MD, USA
Corresponding Author:
Elliot K. Fishman, The Russell H. Morgan Department of Radiology and Radiological Science, Johns Hopkins University School of Medicine, North Wolfe Street, Baltimore, MD 21287, USA.
Email: efishman@jhmi.edu
models. Continued advancements in automated segmentation tools are essential to improve the feasibility and scalability of radiomics-based approaches in routine clinical practice.
Surgical resection of ACC is often complex due to the potential for local invasion with involvement of adjacent organs, or the need for intricate vascular reconstructions. In this context, careful preoperative planning is critical to opti- mize outcomes. Minimally invasive approaches, such as lapa- roscopic adrenalectomy, may be appropriate for selected patients with localized and less aggressive tumour; however, patient selection for these less invasive approaches remains a clinical challenge due to concerns about achieving optimal oncologic control.5 Radiomics and imaging biomarkers offer valuable preoperative assessments of tumour biology, with the potential of individualized risk stratification to improve surgical planning and guiding decision-making. By providing objective, imaging-derived biomarkers, these models could aid in selecting patients who are suitable for laparoscopic resection while avoiding undertreatment in those with more advanced disease that could be better managed through open surgery. Furthermore, accurate preoperative risk stratification can also inform the use of neoadjuvant therapies in initially unresectable patients or patients with high-risk features, and guide decisions regarding adjuvant treatment to improve long-term oncologic outcomes.
There is still much to be done to move these models beyond the development phase and into everyday clinical use. While machine learning and radiomics continue to evolve and meaningful progress is being made toward the integration of imaging biomarkers into clinical workflows, several critical questions remain. Can these models be reliably validated in larger, multi-institutional cohorts? Will advances in computer vision models for tumour segmentation and seamless integra- tion into clinical workflows be sufficient to overcome current barriers to widespread implementation? Ultimately, the field still faces uncertainty regarding translating imaging-derived insights into actionable, real-time tools that meaningfully guide preoperative decision-making and improve patient out- comes. The work by Barat et al represents an important step in this direction, helping us begin to look ahead to a time
when imaging-informed models may genuinely improve out- comes for patients with ACC.
ORCID iDs
Felipe Lopez-Ramirez ID https://orcid.org/0000-0002-1560-9172 Elliot K. Fishman ID https://orcid.org/0000-0002-2567-1658
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was supported by The Lustgarten Foundation. Felipe Lopez- Ramirez and Linda C. Chu receive salary support from The Lustgarten Foundation. Elliot K. Fishman reports grant support from the Lustgarten Foundation, Siemens, and is the co-founder of HipGraphics.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
References
1. Yalon T, Yalon M, Assaf D, et al. Differentiating between adre- nocortical carcinoma and lipid-poor cortical adenoma: a novel cross-sectional imaging-based score. Surgery. 2023;173(1):35- 42. doi:10.1016/j.surg.2022.07.029
2. Ahmed AA, Elmohr MM, Fuentes D, et al. Radiomic mapping model for prediction of Ki-67 expression in adrenocortical carci- noma. Clin Radiol. 2020;75(6):479.e17-479.e22. doi:10.1016/j. crad.2020.01.012
3. Libé R, Borget I, Ronchi CL, et al. Prognostic factors in stage III-IV adrenocortical carcinomas (ACC): an European Network for the Study of Adrenal Tumor (ENSAT) study. Ann Oncol. 2015;26(10):2119-2125. doi:10.1093/annonc/mdv329
4. Barat M, Eltaher M, Moawad AW, et al. A computed tomogra- phy-based score to predict survival in patients with adrenocorti- cal carcinoma: a proof-of-concept study. Can Assoc Radiol J. Published online May 4, 2025. doi:10.1177/08465371251335170
5. Shariq OA, Mckenzie TJ. Adrenocortical carcinoma: current state of the art, ongoing controversies, and future directions in diagnosis and treatment. Ther Adv Chronic Dis. Published online July 20, 2021. doi:10.1177/20406223211033103