Is Lassie the Answer?

In Los Angeles, KABC’s Denise Dador supplies another heartwarming story about olfactory service dogs: “Type 1 diabetes patients get help with blood sugar from scent-detection dogs.”

Let’s stipulate that there are odor cues associated with diabetes (and skin and lung and prostate cancer). Let’s also stipulate that dogs can be trained to recognize these odors. For me, the question remains: Is this a practical and cost-effective way of dealing with the problem?

Dador quotes the anxious mother of a diabetic child: “She could die of a low blood sugar during the night.” So . . . the alert dog never sleeps?

According to Dador, fully trained diabetic alert dogs cost about $20,000. She also notes that about 3 million Americans take insulin for their diabetes. What Dador doesn’t do is the math: a fully implemented diabetic dog solution would cost $60 billion dollars.

Meanwhile, a pair of Italian researchers has just published a mini-review titled “Canine olfactory detection of cancer versus laboratory testing: myth or opportunity?” Giuseppe Lippi and Gianfranco Cervellin reviewed the scientific studies in which dogs were used to detect the scent of bladder, ovary, breast, prostate and skin cancer. While the sensitivity and specificity of disease detection were impressive in some studies, they were less so in others.

In fact, Lippi and Cervellin find that “the most problematic issue” in these studies is “the large heterogeneity of performance” by the canine sniffers. They attribute this variability to differences in dog breeds, training methods, tissue sample preparation, as well as intrinsic smellability of different cancers. Of course, these are hurdles that can be cleared with further research and refinement of methods.

Lippi and Cervellin raise another, more substantial issue: confounding comorbidities.

Another important drawback is that the animals were tested to distinguish between normal and cancer samples (either being cancer tissue, blood or urine), but they have not been tested so far to differentiate cancer patients from those suffering from other comorbidities. A variety of non-cancerous diseases (e.g., those characterized by inflammation, infection, or necrosis) might produce confounding or even overlapping biochemical signals that might confuse the animal, decreasing its detecting performance. At best, additional and more specific training might be required for the animals to distinguish between confounding diseases and cancers.

To me, this fair-minded analysis shows there is still a long way to go before scent detection dogs become a routine part of medical diagnosis and care.

The study discussed here is “Canine olfactory detection of cancer versus laboratory testing: myth or opportunity?” by Giuseppe Lippi and Gianfranco Cervellin, published in Clinical Chemistry and Laboratory Medicine 50:435-439, 2012.