Preoperative biopsy evaluation of chemotherapy-associated liver injuries: looking for a needle in a haystack? Comment on "prospective evaluation of accuracy of liver biopsy findings in the identification of chemotherapy-associated liver injury".

Abstract

W e read with great interest the study by Viganò et al regarding the use of preoperative biopsy to evaluate chemotherapyassociated liver injuries (CALIs). With the increasing use of preoperative chemotherapy, concern has grown that CALIs may be more prevalent and, in turn, adversely affect perioperative outcomes. Our group has previously shown that patients who undergo a major hepatic resection in the setting of steatohepatitis are at risk for increased perioperative mortality. Other studies, however, found no association between simple steatosis or sinusoidal dilatation and outcome. The prospective trial by the European Organisation for the Research and Treatment of Cancer examined the use of perioperative chemotherapy and found a small increase in perioperative complications in the treatment arm but no difference in mortality. Despite data to suggest that the risk of CALIs may be overstated for most patients, surgeons continue to use myriad tests in an attempt to assess preoperative liver function after preoperative chemotherapy. In the present report, Viganò et al assess the accuracy of direct pathological assessment of the liver with needle biopsy findings. The authors report that the overall sensitivity and accuracy of biopsy findings for CALIs is very poor. Based on previous studies, the overall incidence of clinically relevant CALIs has been noted to be relatively low. Viganò et al found that only about one-quarter of patients had any evidence of severe steatosis, sinusoidal dilatation, or steatohepatitis. Therefore, not surprisingly, the authors recommend against routine use of needle biopsy. Although diagnostic accuracy traditionally has not been thought to be directly affected by factors such as the prevalence of disease, some investigators have pointed out that clinical variability may cause sensitivity and specificity to vary with prevalence. For example, a patient population with a higher disease prevalence may include patients with more severe disease; therefore, a test may perform better in this population. As such, although Viganò et al provide compelling data against the routine use of needle biopsy, the study does not conclusively answer perhaps the more relevant clinical question: Should needle biopsy be used selectively in a population with a suspected higher prevalence of CALIs (eg, patients who are obese or diabetic, have metabolic syndrome, or underwent 6-8 cycles of chemotherapy)? In contrast to clinical accuracy, the clinical efficacy of a test refers to the practical value or the utility of a test for a particular clinical situation. As noted by Remaley et al, the 2 factors that have a large effect on clinical efficacy, but not on clinical accuracy, are prevalence and the cost of misclassifications. In considering needle biopsy, the surgeon must consider the implications of misclassifying CALIs (eg, a false-positive or a false-negative result). Would misclassification dramatically alter the therapeutic plan, operative approach, or need for portal vein embolization? Such questions can only be answered based on the particulars of the specific case. We commend the authors and agree with the conclusion that routine preoperative needle biopsy lacks clinical accuracy and should not be recommended. Whether in high-risk populations the selective use of needle biopsy to assess CALIs lacks any clinical efficacy and therefore should be universally abandoned remains a topic of controversy.