Tumour markers with clinically controlled cut‐offs for suspected cancer

• Published in: European journal of clinical investigation Vol. 51; no. 7; pp. e13523 - n/a
• Main Authors: Bosch, Xavier, Molina, Rafael, Marrades, Ramon, Augé, Josep M., Pellicé, Martina, López‐Soto, Alfonso
• Format: Journal Article
• Language: English
• Published: England Blackwell Publishing Ltd 01-07-2021
• Subjects: Abdominal Pain - physiopathology; Aged; alpha-Fetoproteins - metabolism; Antigens; Antigens, Neoplasm - blood; Biomarkers; Biomarkers, Tumor - blood; CA-125 Antigen - blood; CA-19-9 Antigen - blood; Cancer; Carbohydrates; Carcinoembryonic antigen; Carcinoembryonic Antigen - blood; Carcinoma - blood; Carcinoma - diagnosis; Case-Control Studies; combined panels; Confidence intervals; cut‐offs; Diagnosis; diagnostic accuracy; Dyspnea - physiopathology; epithelial cancers; Female; Hematologic Neoplasms - blood; Hematologic Neoplasms - diagnosis; Humans; Keratin-19 - blood; Liver diseases; Lung cancer; Lymphadenopathy - physiopathology; Lymphoma - blood; Lymphoma - diagnosis; Male; Markers; Melanoma - blood; Melanoma - diagnosis; Middle Aged; Mucin-1 - blood; Neoplasms - blood; Neoplasms - diagnosis; Neoplasms - physiopathology; Nervous System Diseases - physiopathology; Pain - physiopathology; Peptide Fragments - blood; Phosphopyruvate hydratase; Phosphopyruvate Hydratase - blood; Prostate; Prostate-Specific Antigen - blood; Recombinant Proteins - blood; Renal failure; Sarcoma - blood; Sarcoma - diagnosis; Sensitivity and Specificity; Serpins - blood; Signs and symptoms; Squamous cell carcinoma; STARD; Thresholds; Tumors; tumour markers; Weight Loss; cut-offs; diagnostic accuracy; STARD; tumour markers; epithelial cancers; combined panels
• ISSN: 0014-2972; 1365-2362
• DOI: 10.1111/eci.13523

Background
Due to insufficient scientific evidence, panels of tumour markers (TMs) are currently not recommended for use in suspected cancer. However, recent well‐designed studies have revealed a potential clinical value in lung cancer. We analysed the diagnostic accuracy of a panel of 11 circulating TMs with clinically controlled thresholds in the differentiation of cancer from nonmalignant diseases.
Methods
We prospectively recruited 4776 consecutive patients presenting with focal or nonspecific symptoms suggestive of cancer who underwent testing for 11 serum TMs before diagnosis was known. The study abided by 2015 STARD guidelines. Tumour markers included, among others, carbohydrate antigen 19‐9, carcinoembryonic antigen, alpha‐fetoprotein, squamous cell carcinoma‐associated antigen, prostate‐specific antigen (males), neuron‐specific enolase, progastrin‐releasing peptide and carbohydrate antigen 125. Thresholds were adjusted for the presence of kidney failure, liver disease, effusions and dermatological disorders. Results showing ≥1 TMs with concentrations above threshold were considered positive.
Results
Benign diseases were diagnosed in 3281 (68.7%) patients and cancer in 1495 (31.3%), with epithelial cancers in 1214 (77% at stage IV). When applying criteria for controlled thresholds, overall specificity was 98%. Overall sensitivity of the panel in epithelial cancers was 72.2%, positive predictive value 93% and negative predictive value 90.5%. The area under the receiver operating characteristic curve was 0.920 (95% confidence interval, 0.902‐0.924).
Conclusions
By using clinically controlled cut‐offs, the combined panel demonstrated an excellent ability to discriminate epithelial cancers from nonmalignant diseases. However, its use in clinical practice would need formal validation through a multicentre controlled trial assessing a panel‐guided strategy vs. standard diagnosis.