Complication rates of ablation therapies for hepatocellular carcinoma: a difficult comparison with an easy solution

We read the recent debate from Taipei (Gut 2005;54:1151–6) on which is the best image guided ablation therapy for hepatocellular carcinoma (HCC), proposed by Lin and colleagues¹ and Huo and colleagues.² This is an interesting topic for hepatologists and gastroenterologists.

The papers by Lin et al stated that in the treatment of HCC, radiofrequency (RFA) was superior to percutaneous ethanol injection (PEI) and percutaneous acetic acid injection (PAI) with respect to local recurrence, overall survival, and cancer free survival, but RFA also caused more major complications. In contrast, Huo et al stressed that PEI may still be the preferred ablation method for small HCC because the complication rate of RFA may be higher than previously assumed. In fact, the rate of death occurring with the two techniques was higher for RFA (0.3–0.5%) than for PEI (0.09%).³,⁴ Moreover, Huo et al reinforced their point of view by emphasising that RFA is affected by a higher mortality (risk ratio of 10.6‐fold) if the death rate is calculated on the basis of number of sessions of treatment.³,⁴ The letter of Huo and colleagues,²demonstrates that lack of standardisation persists in the reporting results of ablation therapies. This is due to the use of non‐uniform terms and different parameters to calculate the rate of complications.

Although the recent publication of the International Working Group on Image‐guided Tumor Ablation⁵ has standardised terminology, thus facilitating communication between investigators and improving comparison of the results of different ultrasound guided treatments, some doubts remain on which denominator should be used for the rate of complications: number of patients, number of sessions, number of ablations, or number of tumours?

As far as mortality is concerned, it is clear that death should be reported on a per patient basis. If instead we indicate the number of sessions as the denominator, as reported by Huo and colleagues,² then the technique which employs a higher number of sessions to achieve ablation obviously has a lower rate of complications, and this is particularly true when we compare PEI with RFA. As regards treatment related morbidity (that is, major (not death) and minor complications, as defined by the SIR classification⁶) it is more appropriate, in our opinion, to divide the number of complications by the number of ablated nodules (that is, cured).

Unlike the number of sessions, ablations, or tumours, the number of ablated nodules represents the main objective for all ablation techniques and only by using this parameter can we truly compare the incidence of complications that occur after RFA, PEI, or other image guided tumour ablation therapies.

In conclusion, we suggest using a sole parameter as the denominator (that is, number of ablated nodules), as the future for prevention and therapy for HCC will rest on comparative prospective studies based on dedicated databases.⁷