Carcinoma of unknown primary (CUP) refers to a malignancy for which the primary site cannot be identified by conventional diagnostic methods. Based on clinicopathological characteristics, CUP patients can be categorized into two prognostic subgroups. The first subgroup is the favorable-prognosis subgroup (15%–20%) (1), whose clinical and pathological features strongly suggest a specific tissue of origin. This subgroup includes patients with: neuroendocrine carcinoma of unknown primary, peritoneal carcinomatosis of serous papillary subtype, women with isolated axillary lymph node metastases, squamous cell carcinoma involving non-supraclavicular cervical nodes, a single metastatic site of unknown primary, and male patients with osteoblastic bone metastases and prostate-specific antigen (PSA) expression. Tumors in the favorable-prognosis subgroup are typically chemosensitive and demonstrate better outcomes when treated with site-specific therapy. The second subgroup is the unfavorable-prognosis subgroup (80%–85%) (2). Due to the inability to identify the primary tumor, this subgroup is usually managed with empiric chemotherapy and has a poor prognosis. With further research into CUP, more favorable-prognosis subtypes within the unfavorable-prognosis subgroup have been identified, including colorectal-type, lung-type, and renal-type carcinomas of unknown primary (3). These subtypes correspond to specific treatment strategies and are associated with improved survival rates. Therefore, one of the major challenges currently is how to better identify the characteristics of more primary tumors within the unfavorable-prognosis subgroup to enable targeted therapy for these patients. Advances in molecular biology have facilitated classification based on epigenetic alterations in different tumors (4), which aids in implementing optimal individualized treatment for CUP patients and is of significant importance for improving therapeutic outcomes.

Herein, we present two cases in which the primary tumor site remained unidentified despite comprehensive clinical, imaging, pathological, and laboratory investigations. Tumor tissue origin detection based on gene expression profiling proposed potential hypotheses regarding the tissue of origin. Subsequent tumor-specific genetic testing and the observed clinical treatment response provided further corroboration for these findings. The possible tumor origins suggested by the tissue-of-origin assay offered valuable reference for clinicians to adjust the patients’ treatment strategies. This facilitated a shift from traditional broad-spectrum chemotherapy to more targeted approaches, including lung adenocarcinoma-specific chemotherapy combined with targeted therapy, and esophageal carcinoma chemotherapy combined with immunotherapy. Consequently, the patients with CUP received more precise treatment, which may potentially improve their prognosis. This holds significant clinical importance for CUP patients.

We present two relatively typical cases of CUP. Following multiple imaging and pathological examinations, and with the assistance of tumor tissue-of-origin gene testing, definitive diagnoses were established, leading to tumor-specific genetic testing and treatment. In Case 1, after a diagnosis of lung adenocarcinoma, genetic testing revealed an EGFR 19Del mutation. This finding, on the one hand, increased the likelihood of the patient benefiting from targeted therapy; on the other hand, it provided confirmatory evidence for the results of the tumor tissue-of-origin test. In Case 2, following lymph node metastasis, a diagnosis of esophageal squamous cell carcinoma was established, and the patient received chemotherapy combined with immunotherapy. The PFS has now reached 12 months, indicating a favorable therapeutic outcome. This result, on the one hand, demonstrates the efficacy of chemotherapy combined with immunotherapy in CUP patients; on the other hand, it also provides confirmatory evidence for the results of the tumor tissue-of-origin test.

The tumor tissue origin gene detection kit employed in this case is an ex vivo diagnostic method based on quantitative fluorescent PCR of tumor mRNA combined with artificial intelligence. By detecting and analyzing the expression of 90 signature genes in paraffin-embedded tumor samples, it calculates a similarity score to determine 21 common tumor types (5) (Table 1). When the similarity score is ≥45.3, the submitted tumor sample is most likely derived from that specific tumor type. Multiple retrospective studies have evaluated the overall accuracy of this product in CUP as ranging from 92.0% to 97.4% (6–9). Real-world studies indicate that this assay can support the provision of cancer type-specific treatment for 82.3% (116/141) of patients (5). A prospective randomized controlled clinical trial, the first of its kind globally focusing on CUP, provided the first international evidence for the efficacy of “site-specific therapy” in CUP. Studies have shown that patients with CUP whose tumor tissue of origin is predicted via genetic testing and who subsequently receive “site-specific therapy” achieve a median progression-free survival of 9.6 months. This is significantly superior to the 6.6 months observed with “traditional empirical chemotherapy” regimens, representing a 32% reduction in the risk of disease progression and a 9-month extension in median overall survival, thereby effectively improving the prognosis for this patient subgroup (10). However, according to the literature, current studies comparing site-specific therapy with empirical chemotherapy exhibit significant limitations. These include issues with patient recruitment (over-inclusion of chemotherapy-resistant tumor types and prolonged recruitment periods), limitations in study design (observational studies and trials with inherent flaws), heterogeneity in CUP classification methods (e.g., epigenetic analysis versus transcriptomic analysis), and non-comparability of treatment plans (11). Consequently, the inferences drawn from many studies are constrained. An evaluation of recently published literature on CUP has proposed the adoption of two comprehensive clinical trial design schemes: one prospective and one pragmatic. Both are suitable for incorporating the latest diagnostic and therapeutic advances to enhance the quality of CUP research and improve outcomes for a greater number of patients (12). It is believed that with the continuous advancement of CUP research, the classification and treatment strategies for CUP will become increasingly well-defined.

With the continuous advancement of precision medicine, targeted therapies based on driver gene mutations have been incorporated into the standard treatment plans for certain primary tumors. For example, EGFR-TKIs treatment has become the first-line therapy for lung adenocarcinoma patients harboring EGFR mutations (13). However, whether relevant driver gene mutations characteristic of known primary tumors persist in CUP, the feasibility of administering targeted therapy, and its efficacy in this context require extensive clinical investigation. Literature reports indicate the presence of driver gene mutations in CUP, such as tumor protein p53 (TP53), Kirsten rat sarcoma viral oncogene homolog (KRAS) mutations (14–16), as well as alterations in ALK, EGFR, RET, FGFR1, and NTRK1. Patients harboring these mutations may potentially benefit from targeted therapies (17). Among six CUP cases with EGFR mutations reported in the literature who received EGFR-TKIs, four patients achieved a PFS ranging from six to eleven months (17–19). Therefore, it is reasonable to hypothesize that the patient in Case 1 would also likely benefit from lung adenocarcinoma-specific chemotherapy combined with EGFR-TKIs. Nonetheless, due to the limited number of reported cases, the benefit of targeted therapy for CUP patients remains unclear and warrants further validation through clinical practice. Importantly, while driver mutations are significant, evidence suggests their implications are tumor type-specific. Consequently, their value in CUP, where the tumor origin is unknown, remains uncertain. For example, the BRAF V600 mutation can occur in melanoma, colorectal cancer, non-small cell lung cancer, and papillary thyroid carcinoma. However, BRAF inhibitors are effective in melanoma and advanced papillary thyroid cancer patients with BRAF V600 mutations but are inactive in colorectal cancer and non-small cell lung cancer patients harboring the same mutation (20). This also represents a potential diagnostic approach for CUP highlighted in this report: First, clarify the tumor origin, and then perform tumor-specific genetic testing. However, whether this approach is applicable in clinical practice still requires extensive clinical research for exploration.

Immunotherapy represents a profound treatment modality for malignant tumors, demonstrating significant efficacy either as monotherapy or in combination with other chemotherapeutic agents. The use of immune checkpoint inhibitors (ICIs) has markedly improved survival rates in various malignancies (21–23). Studies have indicated that patients with CUP may be particularly sensitive to ICIs (24). In CUP, 28% of patients harbor one or more predictive biomarkers for ICIs response. Specifically, these include: programmed death-ligand 1 (PD-L1) expression ≥5% on tumor cells in 22.5% of patients (with 34% showing expression ≥1%), positive PD-L1 expression on lymphocytes in 58.7% of patients, microsatellite instability-high (MSI-H) status in 1.8% of patients, and a tumor mutational burden (TMB) ≥17 mutations per megabase in 11.8% of patients (25). Analysis suggests that a tumor mutational burden >10 mutations per megabase may provide a potential genomic correlate for ICIs response in CUP patients (26). However, these biomarkers require further validation in the CUP patient population. However, it is crucial to acknowledge that CUP patients often cannot access newer and more effective treatments (such as immunotherapy or molecular targeted therapy) because the approvals for most ICIs and targeted agents are disease-specific (27). Therefore, determining the most likely site of origin for CUP patients is urgently needed, as it facilitates the selection of optimal treatment plans and may improve prognosis and survival rates. In this case, the patient with lymph node metastasis was considered to have esophageal squamous cell carcinoma based on tumor origin testing combined with medical history. After treatment with immunotherapy combined with chemotherapy, the patient’s condition remained stable, and PFS has currently reached twelve months. Two reported cases of CUP achieved complete remission after receiving immunotherapy combined with chemotherapy, indicating favorable therapeutic outcomes (28, 29). Immunotherapy combined with chemotherapy may have a good therapeutic effect on CUP patients, but further extensive clinical practice is still required for exploration.

Regrettably, the patient in Case 1 did not persist with the treatment, and the underlying reasons warrant deep consideration. CUP patients often experience significant psychological and financial stress due to the prolonged diagnostic process, where conventional examinations (e.g., ultrasonography, MRI, CT, PET-CT) fail to provide definitive results and traditional pathology cannot yield a clear diagnosis, leading to anxiety about the disease diagnosis and prognosis. Therefore, an efficient and conclusive diagnostic workflow is crucial for CUP patients. An earlier definitive diagnosis enables the timely initiation of targeted therapy. As healthcare professionals facing complex diagnostic scenarios like CUP, in addition to providing adequate communication to alleviate patients’ psychological distress, it is more important to fully utilize professional diagnostic procedures to help patients obtain a definitive diagnosis promptly. This approach maximizes the potential for patients to receive more specialized treatment and achieve a better prognosis. The diagnostic approach presented in this report may provide valuable reference for the diagnosis of patients with CUP.

This report presents the diagnostic and therapeutic processes of two patients with CUP. Based on the results of tumor origin detection, Case 1 underwent lung adenocarcinoma-specific genetic testing, leading to a shift from a broad-spectrum anticancer chemotherapy regimen to tumor-specific chemotherapy combined with precision targeted therapy. Case 2, who was subjected to immunotherapy combined with chemotherapy, achieved a favorable therapeutic response. A primary limitation of this report is that the patient in Case 1 was lost to follow-up after discontinuing treatment, precluding the acquisition of prognostic data under targeted therapy. However, the main objective of this report is to elucidate the critical role of tumor origin tracing and subsequent tumor-specific genetic testing and personalized treatment in the diagnosis and management of CUP, which has been adequately demonstrated.