Nearly 10 years ago, authors of a meta-analysis examining the effectiveness of hepatic arterial infusion (HAI) chemotherapy for patients with unresectable colorectal liver metastases (CRLM) concluded that it was ''the end of an era. '' 1 Although patients treated with HAI achieved an impressive improvement in their tumor response rate (43 vs. 18 %), no benefit was seen in the ultimate patientimportant outcome of overall survival (OS). Despite the provocative title, the role of HAI for patients with unresectable CRLM continued to be explored at a handful of high-volume, specialized centers. Recent reports of excellent long-term outcomes from patients treated at the Memorial Sloan-Kettering Cancer Center (MSKCC) and the University of Pittsburgh Medical Center (UPMC) suggest that HAI may indeed provide important benefit for selected patients with unresectable CRLM, even in an era of better systemic therapy.
2,3In the accompanying article, Zureikat and colleagues from UPMC report long-term outcomes for 40 patients treated in the contemporary era of systemic therapy (2004)(2005)(2006)(2007)(2008)(2009)(2010)(2011)(2012)(2013)(2014) with a combination of HAI and systemic therapy. The comparator group, consisting of 46 patients with liveronly metastases who received systemic therapy alone, is problematic for three reasons. First, the outcomes are surprisingly poor. The median survival time was 15.3 months from diagnosis compared with an OS in the range of 24-28 months for patients treated in other trials with modern systemic therapy regimens. 4,5 The authors hypothesize that the poor OS for these patients may be attributable to a heavy burden of liver metastases. However, it should be noted that none of these patients had documented extrahepatic metastases, so if anything, their outcomes should have been at least as good as those for patients with diffuse colorectal metastases. Second, the patients in the comparator group received treatment at more than 21 network community sites affiliated with UPMC, whereas the patients in the HAI group received treatment at the UPMC flagship oncologic facility. Finally, the primary tumor was resected in all the patients treated with HAI, whereas only 54 % of the patients in the comparator group underwent resection. These differences in baseline characteristics and concurrent therapies make it difficult to compare the outcomes between the treatment groups. Indeed, the survival curves appear to separate early on, less than 6 months after diagnosis, at which point most patients in the HAI group would not even have undergone an operation to implant the pump.Despite these limitations, the outcomes for the 40 patients treated with HAI offer a glimpse into what is possible with this intensive therapy in a modern chemotherapy setting. The median OS for the patients treated with HAI plus systemic therapy was 32.8 months, substantially better than the outcomes expected with any modern systemic therapy regimen. Interestingly, only three patients achieved a response sufficient to allow resect...