Lung cancer remains the most frequently diagnosed malignancy worldwide, with an estimated Lung cancer remains the most frequently diagnosed malignancy worldwide, with an estimated 2.5 million new cases annually and continues to be the leading cause of cancer-related mortality (1). Histologically, it is broadly categorized as small cell lung cancer (SCLC) and non-small cell lung cancer (NSCLC), the latter comprising adenocarcinoma, squamous cell carcinoma and large-cell carcinoma (2). NSCLC accounts for approximately 80–90% of all lung cancer cases. At presentation, only 15% of NSCLC patients are diagnosed at an early stage, 20–25% present with locally advanced disease and the majority, 55–60%, present with metastatic disease (3).

Therapeutic strategies in NSCLC are stage-dependent. In localized and locoregional stages, surgical resection remains the cornerstone of curative treatment in appropriately selected patients (4). Yet, high recurrence rates following surgery alone have prompted the integration of perioperative systemic therapies, including platinum-based chemotherapy, targeted therapies and immunotherapy, all of which have transformed outcomes in recent years (5–9). In locally advanced, non-resectable disease, concurrent chemoradiation followed by consolidation with durvalumab or osimertinib in epidermal growth factor receptor (EGFR)-mutant tumors represents the current standard of care (10, 11). By contrast, in metastatic disease, therapeutic intent is palliative—the aim is to extend survival and improve quality of life. Thus, accurate staging and precise identification of metastatic spread are of paramount clinical relevance.

The adrenal glands represent a frequent site of metastatic involvement in NSCLC, reported in approximately 16% of cases, after bone (34%), lung (32%) and brain (28%) (12). However, adrenal lesions identified during staging present a well-recognized diagnostic dilemma. While many correspond to true metastases, benign counterparts such as adenomas, hemorrhage, granulomatous disease, or primary adrenal tumors are also common. This distinction carries major therapeutic implications, particularly in patients without other extrathoracic disease, where a misclassification may alter staging, eligibility for curative surgery, or systemic treatment strategy.

18F-fluorodeoxyglucose positron emission tomography/computed tomography (18F-FDG PET/CT) is widely employed in staging and has demonstrated high diagnostic accuracy, with reported sensitivity and specificity approaching 90% and 97% for adrenal lesions, respectively (13). Nonetheless, pitfalls exist: false-negative findings may occur in small (< 10 mm), necrotic, indolent adenocarcinoma metastases, or early lesions, whereas false-positive uptake may be observed in lipid-poor adenomas, hemorrhagic nodules, adrenal hyperplasia, infections, or granulomatous disease (14, 15). Indeed, up to 5% of adenomas may demonstrate FDG uptake exceeding hepatic reference values, mimicking metastases (15). Similarly, computed tomography (CT) and magnetic resonance imaging (MRI) provide valuable morphological characterization, yet remain limited by occasional false-positive and false-negative interpretations (16–18). Although histopathology remains the diagnostic gold standard, biopsy is often restricted by procedural risks, technical limitations and clinical feasibility.

This clinical conundrum underscores the need for reliable non-invasive predictors. Integrative assessment combining imaging characteristics, biochemical parameters and tumor staging may offer enhanced diagnostic performance. Against this background, the present study sought to evaluate the predictive roles of SUVmax (Maximum Standardized Uptake Value) on PET/CT, carcinoembryonic antigen (CEA), cancer antigen 125 (CA125), adrenal lesion size and primary tumor T (tumor size and invasion stage) and N (lymph node spread) stage in differentiating benign adrenal lesions from true adrenal metastases in patients with NSCLC. By doing so, we aim to contribute to the refinement of a clinically applicable diagnostic algorithm for risk stratification and therapeutic decision-making.

This study was designed as a retrospective, cross-sectional analysis. Medical records of all patients diagnosed with lung cancer between 2015 and 2025 at our institution were systematically reviewed (n = 2,194). Among these, patients were eligible if they had a confirmed diagnosis of NSCLC and evidence of an adrenal lesion detected on initial PET/CT performed at the time of diagnosis, without any evidence of extrathoracic metastasis.

Adrenal lesions were defined as metastatic under one of the following conditions:

Lesions that demonstrated stability during follow-up were considered benign. Patients with SCLC, extrathoracic metastases or incomplete clinical, imaging and laboratory data were excluded.

A total of 121 patients meeting these criteria were included in the final analysis and their age, sex, primary tumor T and N stage, serum CEA and CA125 levels, adrenal lesion size, anatomical location and SUVmax were recorded.

Between 2015 and 2025, a total of 2,194 patients were diagnosed with lung cancer in our center. Among these, 121 patients meeting the eligibility criteria were included in the final analysis. Of these, 86 patients (71.1%) had adrenal metastases and 35 patients (28.9%) had benign adrenal lesions. The patient selection flowchart is shown below (Fig. 1).

Baseline demographic and clinical characteristics of the included patients are summarized in Table 1.

AC: Adenocarsinoma; CA-125: Cancer antigen 125; CEA: Carcinoembryonic antigen; mm: millimeters ; ng/mL: nanograms per milliliter; NOS: Not Otherwise Specified; SCC: Squamous cell carcinoma; SUVmax: maximum standardized uptake value; U/mL: units per milliliter.

SUVmax, CEA, CA125 and lesion size parameters were compared between patients with adrenal metastases and those with benign lesions. SUVmax (p = 0.000), CEA (p < 0.01), CA125 (p < 0.01) and lesion size (p < 0.01) were found to be significantly higher in patients with adrenal metastases (Table 2).

ROC curve analysis demonstrated that these variables had moderate to high discriminatory capacity for differentiating adrenal metastases from benign adrenal lesions (Fig. 2).

SUVmax showed the highest diagnostic performance, with a cut-off value of 3.05 yielding an AUC of 0.86 (95% CI, 0.78–0.94), corresponding to 89% sensitivity and 69% specificity (Table 3).

The benign lesions and metastasis groups were compared in terms of T and N stage, together with SUVmax, CEA, CA125 levels and lesion size, based on the cut-off values obtained from the ROC analysis (Table 4).

As shown in Table 4, T stage (p = 0.020), N stage (p = 0.002), SUVmax (p < 0.001), CEA (p = 0.013) and CA125 (p = 0.001) demonstrated significant differences between the benign lesions and metastasis groups. In addition, it was observed that only 20% of both metastases and benign lesions were located on the right side, while 80% were left-sided or bilateral. Further stratification revealed that the prevalence of adrenal metastases was markedly higher in patients with advanced T3–T4 tumors and in those with N2–N3 nodal involvement. Notably, only 7 patients (8.1%) with adrenal metastases were in early stage (T1–2, N0–1), whereas the majority (91.9%) were classified as T3–4 or N2–3 (Table 5).

On univariate logistic regression, higher T stage, higher N stage, SUVmax, serum CEA, serum CA125 and larger adrenal lesion size were all significantly associated with adrenal metastasis (Table 6).

CA-125: Cancer antigen 125; CEA: Carcinoembryonic antigen; mm: millimeters; ng/mL: nanograms per milliliter; SUVmax: maximum standardized uptake value; U/mL: units per milliliter

However, in multivariate analysis, N3 stage (OR 12.75; p = 0.043), elevated serum CEA (OR 1.02; p = 0.038) and adrenal lesion size (OR 1.09 per mm increase; p = 0.037) remained as independent predictors of malignant adrenal involvement. The multivariate logistic regression model explained approximately 48.9% of the variance in malignancy risk.

Lung cancer remains the most frequently diagnosed malignancy worldwide and the leading cause of cancer-related death (1). Adrenal involvement is identified in nearly 16% of patients with NSCLC (12). Despite advances in imaging, studies have shown that radiological methods for characterizing adrenal lesions may yield false-negative or false-positive results in up to 10% of cases (13, 17, 18). In our study, we demonstrated that SUVmax, N stage, T stage, CEA, CA125, and lesion size are important parameters in predicting whether an adrenal lesion represents metastasis or a benign process.

SUVmax is a metric that reflects tissue metabolic activity by measuring ^18F-FDG uptake on PET/CT. It is a well-established fact that tumour cells characteristically demonstrate elevated metabolic rates in comparison to normal tissue. Consequently, the presence of elevated SUVmax values is commonly associated with malignant disease (19). In the present cohort, SUVmax values were found to be significantly higher in patients with adrenal metastases compared with those with benign lesions. ROC analysis identified a cut-off of 3.05, which achieved an AUC of 0.86 with high sensitivity (89%) but only moderate specificity (69%). This finding aligns with the conclusions of previous studies in this field. Cho et al. demonstrated that a cut-off of 2.7 discriminated adrenal metastases with an area under the curve (AUC) of 0.94 (20), while Metser et al. described a threshold of 3.01 providing 98.5% sensitivity and 92% specificity (21). In univariate regression analysis, SUVmax was significantly associated with adrenal metastasis; however, it did not remain an independent predictor in the multivariate model, likely due to correlation with lesion size and N stage. When patients were stratified by the 3.05 threshold, 68.6% of those with SUVmax < 3.05 had benign lesions, whereas 88.4% of patients with SUVmax > 3.05 harbored metastases. These results confirm SUVmax as a strong predictive parameter, while also underscoring its limitations: approximately one-third of low-SUVmax lesions were later identified as early metastases or adenocarcinoma metastases with low FDG avidity (22).

CEA is a glycoprotein produced during intrauterine life that plays an important role in cell adhesion (23). While its use as a prognostic biomarker in colorectal cancer is well-documented, elevated CEA levels have also been associated with metastatic disease burden and the number of metastases in patients with lung cancer (24). In the present study, CEA levels were found to be significantly elevated in the metastasis group. ROC curve analysis indicated a cut-off value of 4.45 ng/mL, yielding moderate discriminatory capacity (AUC 0.69). Notwithstanding these limitations, regression analysis confirmed that elevated CEA was significantly associated with adrenal metastasis. This finding suggests that, while CEA alone is not a robust parameter for distinguishing metastasis from benign adrenal lesions, it may offer valuable complementary information when interpreted alongside clinical, radiological, and other biochemical parameters.

CA125 is a glycoprotein that is widely utilised as a prognostic biomarker, particularly in the context of gynaecological malignancies. Additionally, there have been suggestions that CA125 may be associated with metastatic disease in lung cancer (25). Nevertheless, the prognostic significance and diagnostic accuracy of the test for detecting metastasis in NSCLC remain uncertain. In the present study, CA125 levels were found to be elevated in the adrenal metastasis group in comparison with benign lesions. However, ROC curve analysis with a cut-off value of 24.5 U/mL demonstrated only limited sensitivity and specificity, indicating that CA125 has restricted value as a standalone parameter in distinguishing adrenal metastases from benign lesions.

The present study identified N stage as one of the strongest parameters for predicting adrenal metastasis. As demonstrated in Table 2, the distribution of T and N stages indicates that adrenal lesions detected in patients with T1–2 and N0–1 disease are more likely to be benign. Conversely, our logistic regression model demonstrated that adrenal lesions in patients with N3 stage carried a 12-fold higher risk of representing metastasis, which is a particularly striking finding. These results are consistent with those of previous studies suggesting that lymphatic dissemination plays a key role in the development of adrenal metastases in NSCLC (26).

Furthermore, adrenal lesion size was identified as a significant parameter for distinguishing between benign and metastatic lesions. In the present study, adrenal masses were found to be significantly larger in the metastasis group. Concurrent large series have similarly reported that malignant adrenal tumours tend to be larger than benign lesions (27). ROC analysis identified a cut-off value of 14.5 mm, above which the likelihood of metastasis increased approximately threefold. Nevertheless, the diagnostic performance of this threshold was found to be limited, exhibiting modest sensitivity and specificity. Furthermore, 25.6% of adrenal metastases in the present cohort measured less than 14.5 mm, emphasising that small lesions cannot be assumed to be benign. In accordance with the extant literature, the present findings suggest that larger adrenal size is a useful predictor for metastasis, yet small lesions also warrant careful evaluation (28).

The strengths of our study include the selection of a homogeneous cohort restricted to NSCLC histology, thereby providing a more specific and focused sample. Moreover, by excluding patients with extrathoracic metastases and evaluating only cases with adrenal lesions, we were able to investigate adrenal involvement in greater detail. Another salient strength is the integrated evaluation of clinical, biochemical, and imaging parameters, which collectively furnish a more comprehensive evaluation of metastasis risk.

However, it is important to acknowledge several limitations of this study. The retrospective, single-centre design and the relatively limited sample size may restrict the generalizability of the findings. Furthermore, not all adrenal lesions were subject to pathological confirmation, and reliance on radiological follow-up in some cases may have introduced misclassification bias.

In summary, SUVmax, serum CEA, lesion size, and T and N stage were all significantly associated with the presence of adrenal metastases in patients with NSCLC. For patients with suspicious adrenal lesions, the combined assessment of these clinical, biochemical and imaging parameters provides a more reliable means of differentiating metastasis from benign lesions, thereby enhancing diagnostic accuracy and supporting more appropriate treatment planning.