Hepatic Cryotherapy for Cancer: A Reviewand Critique

Background: Wedge or other nonanatomic hepatic resections, performed in an attempt to spare functional parenchyma, often are not accomplished with clear resection margins and may be complicated by hemorrhage from the depth of the resection. Study Design: The current study describes a technique of cryoassisted hepatic resection that allows for controlled resection with well-defined margins. The early experience in managing 16 tumors in 13 patients is reported. Results: A cryoprobe is inserted into the tumor and freezing performed to a predetermined resection margin using ultrasound control. The ice ball, so formed, is then maintained and excised. The management of these 13 patients was associated with one intraoperative and two postoperative complications, including a death of a patient with cirrhosis who had infected ascites and died as a result of hepatic failure. Conclusions: Cryoassisted hepatic resection seems to be safe and allows resection with good tumor clearance and maximal preservation of functional parenchyma.

RESULTS: A cryoprobe is inserted into the tumor and freezing performed to a predetermined resection margin using ultrasound control. The ice ball, so formed, is then maintained and excised. The management of these 13 patients was associated with one intraoperative and two postoperative complications, including a death of a patient with cirrhosis who had infected ascites and died as a result of hepatic failure. CONCLUSIONS: Cryoassisted hepatic resection seems to be safe and allows resection with good tumor clearance and maximal preservation of functional parenchyma. J. Am. Coll

PAPER DISCUSSION
The use ofcold to destroy cancers was described about 150 years and was well understood 20 years ago but advances in ultrasound and cryomachine/probe technology have, however, allowed this to be applied to some deep cancers, including prostate, kidney and the topic of this essay, the liver. Before considering the paper of Polk et al. in detail, perhaps we could briefly review what is known of hepatic cryotherapy, and our own experience.

1) Hepatoma
There is a sizeable experience of cryotherapy for hepatoma in China with approximately 20% five-year survival and no operative mortality. These patients did, however, receive other treatments, making the data difficult to interpret. The principal advantage of cryotherapy in this patient group is the preservation of hepatic parenchyma.

2) Liver Metastases From Colorectal Cancer
This indication has been responsible for the majority of patients treated in the westen world5. My own experience is of 162 separate procedures in 149 patients, almost exclusively in unresectable patients, the few exceptions being mainly patients considered high risk for resection because of cardiovascular disease. Our technique is described elsewhere6. We have seen one thirty-day death (myocardial infarction) and one late death due to hepatic abscess associated with concurrent colectomy at 10 weeks. Limited perioperative morbidity and short hospital stay was seen even in our early experience7. We have not had the devastating cryoshock phenomenon described by the Pittsburgh group but we do not do a double freeze with complete thaw. There is no doubt that a twin freeze thaw has a considerably more destructive cellular effect 9'1 but even our partial twin freeze thaw produces greater evidence of hepatocellular injury and dramatic fall in platelet countTM. We have chosen to only re-freeze the outer 1 cm of a cryolesion, which is the least definitively destroyed. The issue of the utility of regional chemotherapy following cryotherapy is covered elsewhere 13 but as this therapy is now proven to improve survival 4'5 and is reviewed elsewhere16, we believe that, in patients with more than three lessions, it is mandatory.
CEA reduction has been seen in almost all secretors (82%) and the size ofreduction is ofmedian 84% ofthe preoperative value, the percentage reduction has been highly prognostic of survival13. In more than one half of the CEA responders, this was to the normal range and median survival in this group is in excess of 1000 days, which has to be regarded as most encouraging in patients with multiple (median 3.9) unresectable hepatic metastases17.
The Boston group have mainly treated patients with operable lesions. Of 34 patients treated by cryotherapy alone, there were 1-3 lesions (mean 1.4) 5'18 and, of the first 49 patients treated in Boston, 35 were considered to have no gross residual disease after cryotherapy. Of these, 13 have died, although three deaths were from non-malignant casuses and, of the remaining 22, only 36% appear tumor free at 3-88 months (median 12). I do not yet believe that this follow-up is adequate for cryotherapy to be advocated in patients with operable lesions who have no medical risk factors for resection. I also believe it is premature to mount controlled trails to address this issue. Liver resection can achieve fiveyear survival in almost a half of our patients and with a mortality ofwell under 5%, so what does cryotherapy have to offer to this group? Even when five-year survival of larger numbers of patients with operable disease treated by cryotherapy is known, will cryotherapy offer any significant reduction in risk of the procedure: in skilled hands, I doubt it.
3) Neuroendocrine Although a rare tumour, the role of cryotherapy for neuroendocrine liver metastases is perhaps most exciting and has involved some extensive and time-consuming surgery. We have to date treated without mortality six patients with unresectable hepatic metastases from neuroendocrine tumor. Not only have we seen excellent clinical, radiologic and tumor marker results 9 but the length of benefit in most of this group of patients seems to be considerable. 4) Adjunct to liver resection: edge freeze/ contralateral disease We have used cryotherapy in a different way than the featured paper to treat involved or inadequate hepatic resection margins, where disease free survival is only 20% of the former group at year2, and to treat contralateral disease following lobectomy 21 with to date an apparently lower risk of edge recurrence than might be expected. Only two of twelve edge freeze patients have recurrence at a median of 14 months and, whilst the data is less mature, at least some of the contralobe freeze patients remain tumor free.

5) Future
Hepatic cryotherapy has already been done under laparoscopic guidance 22 but, although we have failed to produce gas embolism in a rabbit model23, I remain concerned about the efficacy and safety ofthis (it really bleeds, sometimes) and also the ease ofplacement ofan arterial perfusion device.
Cryotherapy does offer the ability to destroy large volumes of liver with relative safety. Our data suggests that the adequacy of tumor ablation, as measured by CEA reduction, does affect survival. Indeed, the large group where CEA returns to normal appear (at least as far as three years) to have similar survival prospects to patients who do have resectable disease. Cryotherapy has been a very useful adjunct to liver resection in our hands, although we have used it in a very different manner than Blumgart's group.

CRITIQUE
The use ofcryotherapy to ensure an adequate resection margin is hard for me to understand. Cryotherapy does take time, the equipment is expensive and it does have intra-operative hazards of its own, even if some of the postoperative sequelae of cryotherapy are avoided by this iceballectomy. Clearly, the concept depends upon the ultrasonographic skill of the operator because one will be making ajudgement ofthe cm margin between lesion edge and iceball edge on ultrasound and, of course, lesion edge disappears as soon as it is enveloped by iceball. It must be understood that this cm margin needs to be on every margin of the tumor. The paper is also largely devoid of meaningful data, in that there is no effort to describe disease free recurrence. It is particularly disappointing that there is not even a measure ofthe actual resection margins achieved, i.e., minimum margin in each lesion. Perhaps it would be valuable if compared with some retrospective data for wedges. The contemporary condemnation of wedges in favour of segmentally based resection is also based upon the poor blood loss record of wedges. It seems unlikely that the described technique will alter this and there is not a comprehensive description of the blood loss/transfusion requirements of these patients in this paper.
The cryoablative role of hepatic cryotherapy may well require more data to convince us but its role as a handle or a fender bar would seem less likely to be of importance, but I've been wrong before.