Adjuvant Therapy for Gastric Carcinoma: Closing out the Century

Gastric carcinoma is a discouraging disease. Although we can clearly identify patient- and tumor-related variables that predict outcome, the only reproducible treatment- related variable associated with an improvement in survival is a complete (R0) resection.[1]

Results of two recent prospective, randomized trials suggest that it is the biology of the tumor, rather than extent of surgical resection, that governs patient outcome.[2,3] A detailed analysis of this conclusion is beyond the scope of this review. Suffice it to say that these trial results create the impression that the outcome of patients who present with gastric cancer is predetermined largely by stage at presentation and that, beyond complete surgical resection, little else has been consistently shown to alter the all-too-frequent, inexorable progression of this disease.

In this review, Yao and colleagues present a superb summary of the state of adjuvant therapy for gastric carcinoma, citing results of both historical and recent, national and international, single-center and cooperative group trials. Results of smaller positive trials have not been replicated, and results of most larger trials have shown no advantage of adjuvant systemic therapy. Yao and colleagues conclude that the available data have failed to justify the routine administration of systemic therapy following complete resection of gastric cancer. Their conclusion is supported by a meta-analysis by Hermans et al on the same topic.[4]

Trials of Adjuvant Immunochemotherapy

Two prospective, randomized trials of adjuvant immunochemotherapy not mentioned by the authors, both from the Seoul National University Hospital in Korea, are worthy of review.[5] In the first trial, 73 patients were randomized, after curative resection, to receive postoperative adjuvant immunochemotherapy with fluorouracil(Drug information on fluorouracil) (5-FU), mitomycin(Drug information on mitomycin) (Mutamycin), cytarabine(Drug information on cytarabine), and OK-432 (picibanil; a nonspecific Streptococcus pyogenes immunostimulant). These patients had a survival rate of 45%, which was statistically superior to the rate of 23% in 64 patients followed by observation alone.

In the second trial, 159 patients also received postoperative immunochemotherapy with 5-FU, mitomycin, and OK-432. Their 45% survival rate was statistically superior to the 30% rate observed in 77 patients receiving 5-FU and mitomycin without immunotherapy and to the 24% rate noted in 94 patients followed by observation alone.

Taken together, the results of these two trials suggest that immune stimulation by OK-432 should be an important component of any postoperative (adjuvant) systemic strategy. The two trials appear to have been performed in as uniform a clinical setting as possible. As provocative as the results are, however, they clearly require confirmation in a broader western setting before this approach can be routinely recommended as standard therapy.

Preoperative, Intraoperative, and Molecular Staging

The authors’ comments on the need for careful staging prior to treatment, either curative resection or adjuvant systemic therapy, are well taken. In our institution, we rely heavily on preoperative assessment with laparoscopy and laparoscopic ultrasound.[6] Using these modalities, we have discovered subradiologic disease in up to 33% of patients deemed “resectable for cure” by high-resolution computed tomography (CT). A portion of patients upstaged by laparoscopy were found to have very low-volume peritoneal implants that could well have been missed, even by routine exploratory laparotomy.

In a study by Keighley and colleagues, routine intraoperative staging by biopsy of portal or preaortic lymph nodes revealed clinically unsuspected metastatic disease in 19 (22%) of 84 patients undergoing “curative” gastrectomy.[7] Furthermore, these investigators found that patients with positive N4 nodes fared very poorly, with a median survival of only 4.5 months, comparable to that of patients undergoing palliative resection. The design of the prospective trial conducted by the Dutch Gastric Cancer Group wisely accommodated this observation by insisting on a histologically negative N3 (periaortic) node biopsy prior to randomization of patients to either a D1 or D2 lymph node dissection.

Finally, in addition to radiologic and pathologic staging, we have now definitely entered the era of molecular staging. Yao et al allude to studies exploring the prognostic value of p53 mutations in gastric cancer cells. Other investigators have assessed the prognostic value of detecting circulating tumor cells in the peripheral blood or bone marrow by immunohistochemistry[8] or the reverse transcriptase polymerase chain reaction.[9] This extremely interesting avenue of research (demonstrated to be of value in other tumors) holds great promise for differentiating good-risk from poor-risk patients with gastrointestinal malignancies.

Designing the Optimal Trial of Adjuvant Therapy

It is unlikely that a truly optimal trial of adjuvant therapy will ever be carried out, as even very minor maldistributions of adverse prognostic factors can either obscure a clinically relevant result (false-negative trial) or masquerade as an unreproducible clinical effect of treatment (false-positive trial). It is imperative that patients being enrolled in any such trial be comprehensively staged both preoperatively and intraoperatively, so that the population be as uniform as possible prior to the initiation of adjuvant therapy.

Surgery for gastric cancer is relatively complex, with innumerable small intraoperative decisions contributing to the final procedure. Many of these decisions are made on the basis of surgical expertise and judgment, and are impossible to categorize and quantify completely. Other variables, including the objective estimates of extent of lymph node dissection, are very difficult to standardize, even in the context of a trial in which participating surgeons are actively proctored and monitored.[10] In addition, the importance of a consistent, thorough pathologic evaluation of the surgical specimens, with well-documented stage migration of patients after more thorough evaluation of dissected nodes, has been well documented in comparisons of Japanese vs western surgical series.[11]

All of the above factors must be well balanced before another difficult decision is made, namely, the design of the optimal adjuvant regimen. Specific issues that need to be addressed include the following: Which drugs and/or biological agents should be used, and in what doses, sequence, and duration? Should adjuvant therapy be administered preoperatively or postoperatively? Should adjuvant chemotherapy be given with or without regional modalities, such as radiation or intraperitoneal therapy?

No ‘One-Size-Fits-All’ Approach

Finally, it is likely that gastric cancer is not one disease, but rather a spectrum of diseases. As such, it may be that no single adjuvant therapy is appropriate for all patients. I suspect that the future in managing these patients lies not in the routine administration of the same modestly effective treatment to all patients after complete resection, but rather, in the tailoring of treatment recommendations to the characteristics of the individual patient’s tumor.

The work of Metzger, Dannenberg, and colleagues, mentioned by the authors, represents an enormously important first step in this direction.[12] By defining the molecular profile of individual tumors, these researchers have begun to identify a small group of patients whose tumors are much more likely to respond to conventional chemotherapeutic agents. Equally important, they have begun to define a larger group of patients who are unlikely to benefit from conventional therapies and for whom other novel approaches should be sought.