In a study published in the July 1st Journal of Clinical Oncology, Indian investigators showed that dogs can detect cancer with a sensitivity of 90.8% and a specificity of 91.3%. (Canine Olfaction Combined with Bayesian Modeling for Multicancer Detection from Breath Samples: A Phase 2 study in India, Kulgod S, J Clin Oncol July 1, 2026).
In the Indian study of over 3200 subjects, trained dogs were used to detect cancer by smelling the breath of patients with biopsy proven cancers. Controls were all healthy volunteers.
The receiver operator curve (ROC) for this technology, a critical measure of clinical accuracy, of 0.962 is better than virtually all cancer tests currently being developed for multicancer early detection (MCED).
Indeed, with the recent very disturbing failure of a highly touted blood DNA test that was designed to detect early cancers that was conducted in over 140,000 English subjects, these canine findings take on a very new significance.
What is evident from the Indian study is that they are measuring cancer metabolism not cancer genetics.
As we have shown in our published series, cancer utilizes nutrients and metabolism differently from normal tissues. This cancer metabolic byproducts show up in the patients exhaled breath as volatile, organic compounds (VOC).
In our publications in breast cancer, ovarian cancer, pancreatic cancer and other diseases, we have established that metabolic measures in each patient’s blood can identify cancer in its earliest stages.
Dog olfaction is between 10,000 and 100,000 times more sensitive than humans. Indeed, dogs can detect chemicals at 1 part per trillion, while humans are only 1 part per million. This level of detection (LOD) puts the dog’s nose right smack in the middle of the mass spectrometry platform that we apply in our laboratory.
The implications are several: First, metabolic measures are more sensitive than genomic measures for early cancer detection. Second, the metabolic basis of cancer is confirmed, establishing that biochemical changes are highly accurate measures for cancer diagnosis, prognosis and outcome. Finally, utilizing modern mass spectrometry, we may not only detect cancer, but identify the specific metabolites associated with the malignant transformation that we have shown can correlate with diagnosis, prognosis and survival.
The future of cancer medicine will not be genomic but instead, metabolomic.