Endoscopy 2021; 53(07): 737-738
DOI: 10.1055/a-1312-6330

Bile volatile organic compounds, smelling trouble

Referring to Navaneethan U et al. p. 732–736
Enrique de-Madaria
1  Gastroenterology Department, Alicante University General Hospital, ISABIAL, Alicante, Spain
2  Department of Clinical Medicine, Faculty of Medicine, Miguel Hernández University, Elche, Spain
› Author Affiliations

The sense of smell, or the ability to detect volatile chemicals, is conferred by olfactory receptors in the nasal cavity. It helps us detect both dangers (e. g. spoiled food, poisons, hostile creatures) and opportunities (e. g. delicious food, a good Ribera del Duero wine). Physicians have used the sense of smell to assist in the differential diagnosis of disease since Hippocratic times, and gastroenterologists are accustomed to detecting the specific smells of liver disease (i. e. fetor hepaticus), melena, and other conditions. Dogs can be trained to detect scents on the breath and in stool samples indicative of colorectal cancer (CRC) [1] and even COVID-19 [2].

“... volatile organic compound analysis was highly accurate in distinguishing between patients with pancreatic cancer and those with chronic pancreatitis, but it remains unclear whether the model detected a direct effect of tumor metabolism on the bile or whether confounders played a role.”

Volatile organic compounds (VOCs) are a diverse group of carbon-based chemicals that are volatile at ambient temperature [3]. Can we use an electronic nose, an objective chemical analyzer, to diagnose diseases accurately? Will VOC biomarkers be helpful in real clinical practice in the future? Are we advancing to the use of gas biopsy for detection of inflammation or cancer? Specific breath VOC patterns described in ulcerative colitis could differentiate active from inactive disease, and patients with ulcerative colitis from controls [4]. Blood VOCs might also be helpful, as suggested in a study that compared CRC patients with controls [5]. Specific breath VOCs have also been described in patients with CRC [6], with an overall accuracy of 76 %.

In this issue of Endoscopy, Navaneethan et al. performed a very interesting study based on a prospectively maintained database that included clinical data and bile samples from patients undergoing endoscopic retrograde cholangiopancreatography (ERCP) [7]. The study involved a discovery cohort of 57 patients (19 % with chronic pancreatitis and 81 % with pancreatic cancer) and a validation cohort of 31 patients (39 % chronic pancreatitis, 61 % pancreatic cancer). In the study, 22 bile VOCs were analyzed with mass spectrometry by a blinded investigator. Some VOCs were associated with pancreatic cancer or chronic pancreatitis, and a diagnostic model that combined ammonia, acetonitrile, and trimethylamine (TMA) showed 93.5 % sensitivity and 100 % specificity, with a positive predictive value of 100 % in the discovery cohort. Moreover, the model showed 100 % sensitivity and specificity in the validation cohort.

On one hand, these results are encouraging; the VOC analysis was highly accurate in distinguishing between patients with pancreatic cancer and those with chronic pancreatitis in both the discovery and validation cohorts. On the other hand, it remains unclear whether the model detected a direct effect of tumor metabolism on the bile or whether confounders played a role. The median total bilirubin values were 8.35 mg/dL in the pancreatic cancer group vs. 0.8 mg/dL in the chronic pancreatitis group in the derivation cohort, and 4.3 vs. 0.65 mg/dL, respectively, in the validation cohort. This difference reflected differences in clinical practice: patients with pancreatic cancer undergo ERCP to relieve biliary obstruction, whereas patients with chronic pancreatitis usually undergo ERCP to relieve pancreatic duct obstruction. Does aged bile smell different? We can guess that differences in bile composition, due to stasis and bacterial contamination (TMA may be synthesized by anaerobic gut bacteria [7] [8]), might result in specific VOC patterns. Furthermore, compared with pancreatic cancer patients, most patients with chronic pancreatitis were younger, and more than half were active smokers (vs. 13.5 % in the cancer group). Moreover, alcohol is the most frequent etiology of chronic pancreatitis, but it plays a less important role in pancreatic cancer. Alcohol and tobacco might induce particular VOC patterns; for example, elevated concentrations of acetonitrile, a component of cigarette smoke, were found in patients with chronic pancreatitis [7]. Finally, all patients with pancreatic cancer were in the advanced stages of disease (III or IV), where it is easier to make a differential diagnosis between pancreatic cancer and chronic pancreatitis, compared with examining a resectable pancreatic mass without extrapancreatic involvement.

To conclude, VOC analysis is a promising tool for differentiating pancreatic cancer from chronic pancreatitis, but confounders should be controlled. Future studies should look for specific VOC patterns of pancreatic cancer in patients with resectable disease, irrespective of bile duct obstruction, smoking, and alcohol consumption. Furthermore, looking for specific VOCs in the breath, stool, or blood can avoid the invasiveness of ERCP. This endeavor will require multicenter, and probably international, involvement; however, a differential diagnosis for an aggressive disease such as pancreatic cancer clearly deserves the effort.

Publication History

Publication Date:
24 June 2021 (online)

© 2021. Thieme. All rights reserved.

Georg Thieme Verlag KG
Rüdigerstraße 14, 70469 Stuttgart, Germany