Inaccurate preoperative biopsy findings could mean that patients who are candidates for endoscopic resection (ER) are unnecessarily undergoing esophagectomy, a procedure with greater risks of morbidity and mortality, reported Ravi S. Shah, MD, of Cleveland Clinic, and colleagues.
Cleveland Clinic
Dr. Ravi S. Shah
“It is unclear how accurate tumor differentiation on endoscopic biopsies is and if it can be used for clinical decision-making,” the investigators wrote in Techniques and Innovations in Gastrointestinal Endoscopy. “Given that tumors may be considerably heterogeneous in gland formation, the limited amount of tissue obtained from endoscopic forceps biopsies may not be representative of the entire tumor for pathologic grading, which may result in discrepant tumor grading between biopsy and resection specimens.”
While previous studies have compared esophagogastroduodenoscopy-guided biopsy results with histological findings after surgical resection, scant evidence is available to compare biopsy findings with both surgically and endoscopically resected tissue.
Despite this potential knowledge gap, “many patients with poorly differentiated EAC on preresection biopsy do not undergo ER, with the belief that the final resection pathology would be noncurative,” the investigators noted.
To help clarify how congruent pre- and postoperative biopsies are for both resection methods, Dr. Shah and colleagues conducted a retrospective study involving 346 EAC lesions. Samples were drawn from 121 ERs and 225 esophagectomies performed at two tertiary referral centers. Preoperative and postoperative findings were compared for accuracy and for level of agreement via Gwet’s AC2 interrater analysis.
For all evaluable lesions, preoperative biopsy had an accuracy of 68%, with a “substantial” agreement coefficient (Gwet’s AC2, 0.70; P less than .001). Accuracy in the esophagectomy group was similar, at 72%, again with “substantial” agreement (Gwet’s AC2, 0.74; P less than .001). For the ER group, however, accuracy was just 56%, with a “moderate” level of agreement (Gwet’s AC2, 0.60; P less than .001).
“We speculate that the discrepancy of tumor differentiation on endoscopic forceps biopsies and resection specimens is due to nonrepresentative sampling of tumors to accurately determine the percentage of gland formation and thus tumor grade,” the investigators noted.
Further analysis showed that 22.7% of moderately differentiated tumors were upgraded to poorly differentiated upon final histology. Conversely, 19.6% of poorly differentiated tumors were downgraded to moderately differentiated. Downgrading was even more common among T1a tumors, 40% of which were changed from poorly to moderately differentiated between pre- and postprocedural histology.
These latter findings concerning downgrading are particularly relevant for clinical decision-making, the investigators noted, as patients with poorly differentiated EAC on preoperative biopsy are typically sent for esophagectomy – a more invasive and riskier procedure – out of concern that ER will be noncurative.
“If poor differentiation was the only high-risk feature, these patients may have unnecessarily undergone esophagectomy,” Dr. Shah and colleagues wrote. “Especially in marginal surgical candidates, staging ER should be considered in patients with early esophageal cancer with preoperative biopsies showing poorly differentiated cancer.”
The investigators disclosed relationships with Medtronic, Lucid Diagnostics, Lumendi, and others.
