CA 19-9 remains the most commonly prescribed serum tumor marker in digestive oncology, yet its positioning in the diagnostic decision tree is increasingly questioned. Its sensitivity and specificity, often cited around eighty percent in symptomatic cohorts, mask structural flaws that daily clinical practice highlights.
False negatives of CA 19-9 in Lewis negative patients
The antigenic determinant of CA 19-9 is a sialylated derivative of the Lewis a blood group pentasaccharide. Individuals with Lewis phenotype (a- b-) are biologically unable to synthesize this marker. Therefore, we observe undetectable levels in these patients, even in the presence of advanced pancreatic adenocarcinoma.
This limitation is not anecdotal. The proportion of Lewis negative subjects varies across populations, and a normal CA 19-9 never excludes pancreatic cancer. Any digestive oncologist who prescribes this test without checking Lewis status risks producing a falsely reassuring result.
In practice, Lewis phenotyping is rarely performed routinely. The clinical reflex is rather to never base a therapeutic decision on an isolated CA 19-9, regardless of its level. The oncologists’ opinion on the CA 19-9 antigen indeed converges towards this systematic caution in the face of negative results.
CA 19-9 and CRP: eliminating the inflammatory pathway before the tumor pathway
A moderately elevated CA 19-9 accompanied by an elevated C-reactive protein (CRP) increasingly directs oncologists towards a priority search for non-tumoral causes. Cholangitis, chronic pancreatitis, benign biliary obstruction, cirrhosis: these conditions can cause sometimes significant elevations of the marker.
This CRP/CA 19-9 crossover reflex is recent in decision-making algorithms. Classic technical sheets on the marker mention benign causes of elevation but do not propose a concrete diagnostic sequence. Specialized centers in digestive oncology now integrate this crossover as a systematic step:
- An elevated CA 19-9 with normal CRP reinforces tumor suspicion and justifies rapid complementary imaging
- An elevated CA 19-9 with elevated CRP first necessitates exploration of inflammatory or infectious biliary causes
- An isolated elevated CA 19-9, without clinical context or suspicious imaging, is insufficient to initiate a heavy oncological workup
This approach avoids anxiety-inducing diagnostic cascades for the patient and unnecessary costs.
Preoperative prognostic value of CA 19-9 in pancreatic surgery
It is probably in prognostic stratification that CA 19-9 retains its greatest clinical utility. A preoperative level within normal values is associated with significantly prolonged median survival compared to patients with elevated levels before resection.
We use this data to refine the discussion in multidisciplinary consultation meetings. A very high CA 19-9 before surgery, especially if it does not decrease after biliary drainage, may steer towards neoadjuvant chemotherapy rather than immediate resection.
Postoperative kinetics and follow-up under treatment
The absolute value of CA 19-9 at a given moment matters less than its kinetics. A progressive decrease under chemotherapy correlates with a better therapeutic response, while a rise often precedes radiological confirmation of progression.
However, we recommend not modifying a chemotherapy protocol solely based on a variation in CA 19-9. Scheduled imaging (thoraco-abdominopelvic CT scan, liver MRI) remains the final arbiter. The marker serves as a warning signal, not an autonomous decision-making criterion.
CA 19-9 and detection of recurrences: a secondary tool
Post-resection pancreatic surveillance protocols assign a limited role to CA 19-9 in detecting early recurrences. The marker may remain normal despite a locoregional recurrence confirmed by imaging. This finding, documented in several hospital care pathways, explains why surveillance primarily relies on a sequential imaging program.
An increasing CA 19-9 between two scheduled scans may justify advancing the next examination. In contrast, a stable or low level in no way exempts from the imaging scheduled in the follow-up calendar.
Limitations in biliary and gastric cancers
CA 19-9 is also prescribed in cholangiocarcinoma and certain gastric cancers, but its diagnostic performance is even more variable there. The heterogeneity of decision thresholds from one assay kit to another complicates the comparison of results between laboratories. Inter-technical variations, related to the antibody used and the absence of an international reference standard, remain an unresolved issue.
- A result obtained by electrochemiluminescence is not directly comparable to a measurement by fluorimetry
- The freezing or dilution of a serum sample can reveal masked epitopes and falsely elevate the result
- Longitudinal follow-up of a patient requires using the same assay technique in the same laboratory
Francophone decision-making algorithms and the real place of CA 19-9
Francophone digestive oncology algorithms explicitly classify CA 19-9 as an aid for monitoring, not a tool for early detection. Its positive predictive value for screening in the general population is too low to justify screening use, even in high-risk cohorts (family history, hereditary pancreatitis, BRCA2 mutation).
This position does not diminish the marker’s utility. It refocuses it on three specific functions: preoperative prognostic stratification, monitoring treatment response, and serving as an alert signal between two follow-up imaging sessions. Outside of these three contexts, prescribing a CA 19-9 exposes one more to diagnostic confusion than to real clinical benefit.
CA 19-9 is neither obsolete nor sufficient. Its interpretation requires systematic cross-referencing with imaging, inflammatory context, and the patient’s Lewis phenotype. An oncologist who reads this test in isolation risks underdiagnosing or over-investigating, two pitfalls that current recommendations specifically seek to avoid.