I agree with Dr. Cho that radiofrequency ablation (RFA) and surgical resection each offer certain advantages and are also each associated with certain shortcomings. In fact, the main objective of our article was not to directly compare the outcomes of resection and RFA for patients with hepatocellular carcinoma (HCC) ≤2 cm. This is better done by large cohort studies such as the one recently published by Hasegawa et al. Rather, the main aim was to report the outcomes achieved with resection at two large-volume Western centers, thereby providing some context by which to judge the relative role of resection for treatment of very early HCC. A secondary objective was to demonstrate the heterogeneity that exists even among such early tumors and determine which characteristics were associated with survival and recurrence.
I would like to address the points brought up by Dr. Cho:
- Cirrhosis is a pathological diagnosis. The study by Livraghi et al. did not include any biopsies of the nontumoral liver. Thus, while all of their patients may have had liver disease, it is impossible to state that they were all cirrhotic.
- As stated in the Materials and Methods section, Mount Sinai offered resection to patients with portal hypertension until 2002. Milan continues to offer limited resection in patients with portal hypertension. In fact, 25 (19%) of the patients in our study had a platelet count below 100,000/μL, which most centers would define as having significant portal hypertension.
- A cutoff of 1.5 mg/dL for bilirubin is somewhat arbitrary and the reference provided by Dr. Cho uses a cutoff of 1 mg/dL. Nevertheless, our study included 54 (41%) patients with bilirubin >1 mg/dL and 19 (15%) with bilirubin >1.5 mg/dL. Bilirubin with a cutoff of 1 mg/dL was not significantly correlated with survival even on univariate analysis in our study (Supporting data). Likewise, bilirubin with a cutoff of 1.5 mg/dL was also not correlated with survival (unpublished data).
- The 5-year survival for the entire cohort in the study by Livraghi et al. was 55%. It is only the patients who were classified as “operable,” defined as age <75 years, alanine aminotransferase (ALT) <3× normal, bilirubin <1.5 mg/dL, AND absence of portal hypertension, who achieved a 5-year survival of 68%.
The mean size of the tumors in the study by Livraghi et al. is never reported. While our study included patients with arterial enhancement >2 cm, none of the tumors were larger than 2 cm on pathological examination, which is the gold standard for assessing tumor size. Like the true size of the tumors treated, the incidence of vascular invasion in the Livraghi et al. study will also never be known due to lack of pathological data. We found that the incidence of vascular invasion was only marginally lower in patients with tumors ≤15 mm (22%) compared to those with tumors 16-20 mm (29%) (P = NS, unpublished data).
Nevertheless, Dr. Cho may be correct in that the Livraghi et al. study treated significantly smaller tumors with a lower incidence of vascular invasion using RFA in order to achieve results similar to what we obtained by resecting larger tumors with a higher incidence of vascular invasion. Due to the lack of pathological data, we will never know.
In conclusion, I agree with Dr. Cho that direct comparisons should not be made between our study and the study by Livraghi et al. These types of discussions highlight one of the fundamental advantages of resection over RFA, pathological data from the resected tumor and the surrounding liver, both of which have been shown to have a significant impact on prognosis. Because there is such a wide range of heterogeneity, even among such early tumors, clinicians cannot truly know what it is that they have treated without tissue.
Sasan Roayaie, M.D.
Department of Surgery, Mount Sinai Medical Center, New York, NY