Many patients who have undergone optimal cytoreductive surgery and subsequent therapy for epithelial ovarian cancer will have no evidence of disease at the completion of treatment. Tumor markers and radiologic assessments have proven to be too insensitive to exclude accurately the presence of subclinical disease. Therefore, a common technique used to evaluate residual disease has been the second-look operation. 279-281 A second-look operation is one performed on a patient who has no clinical evidence of disease after a prescribed course of chemotherapy, in order to determine the response to therapy. Most often, patients have undergone a formal reassessment laparotomy. The laparoscope also has been used in some of these cases but there is a 35% false-negative rate if laparoscopy is used as a second-look procedure.

If a second-look operation is performed, the proper surgery should be carried out to maximize the likelihood of detecting any microscopic residual disease. The technique of the second-look laparotomy is essentially identical to that of the staging laparotomy. The operation should be performed through a vertical abdominal incision. The incision may be initiated below the level of the umbilicus, so that if pelvic disease is detected in the absence of any palpable upper abdominal disease, a smaller incision might suffice. The incision can be extended cranially as needed. After first obtaining multiple cytologies, biopsies of the peritoneal surfaces should be performed, particularly from any areas of previously documented tumor. These are the most important areas to biopsy because they are most likely to contain residual cancer. Any adhesions or surface irregularities should be sampled. A pelvic and periaortic lymph node dissection below the level of the inferior mesenteric artery should be carried out in those patients whose nodal tissues were not previously removed, which is the case for most patients with stage III disease.

The ability to obtain a complete remission with a platinum-based therapy is dependent on the extent of residual disease at the time chemotherapy is initiated. Clinical trials have uniformly demonstrated that patients who have bulky disease (any tumor nodule greater than 2 to 3 cm in diameter) have less than a 10% likelihood of achieving a complete remission with induction chemotherapy with a platinum-based regimen. In contrast, patients who have small-volume disease after the initial laparotomy have a three to four times higher likelihood of achieving a complete remission. There is no evidence that prolonging the number of cycles of induction chemotherapy will increase the complete remission rate. In a prospective randomized trial of five versus 10 cycles of the PAC regimen, there was no statistically significant difference in overall survival and only 1 of 12 patients who were not in complete remission after five cycles of PAC achieved a complete remission with an additional five cycles of therapy. It appears that the maximum benefit from a platinum-based regimen is obtained with five to six cycles of chemotherapy.

The levels of CA125 have been correlated with the findings at second-look surgery. Positive titers are useful in predicting the presence of disease, but negative titers are an insensitive marker for the absence of disease. In studies performed to date, the predictive value of a positive test has been 96%, that is, if the level of CA125 was positive (> 35 U/mL), disease was almost always detectable in patients at the second-look. In one such prospective analysis, the predictive value of a negative test was only 56%, that is, if the level was less than 35 U/mL, disease was present in 44% of the patients at the time of the second-look. Therefore, the CA125 is not sensitive enough to exclude subclinical disease in many patients. Serum CA125 titers can be used to follow those patients during chemotherapy whose titer was positive at the initiation of therapy. Changes in antigen level (ie, falling, rising, or plateauing) generally correlates with response. Those patients with persistently elevated levels of CA125 after 3 months of treatment most likely have resistant tumors. Persistently elevated or rising titers on treatment usually indicate treatment failure and suggest that continuation of the current regimen is unlikely to be of value.

Because second-look laparotomies have not been shown to influence patient survival, they should be done only in a research setting, as second-line or salvage therapies in patients with persistent cancer have not yet demonstrated a clear improvement in the overall survival. The findings at a second-look procedure do, however, correlate with subsequent outcome and survival. Patients who have no histologic evidence of disease have a significantly longer survival compared with those who have microscopic or macroscopic disease documented at laparotomy. About 50% to 70% of patients with negative second look examinations remain free of disease at 5 years, whereas the median survival of patients with any disease (microscopic and macroscopic) is 12 to 18 months. The extent of residual disease documented at the second-look procedure also correlates with patient survival. Patients with microscopic disease only have a 5-year survival of 40% to 50% (median of more than 36 months), compared with 12 months for patients with any evidence of macroscopic disease. Clearly, it is not possible to sample every potential site of disease. In addition, disease can become clinically apparent at sites that are impossible to examine, such as the liver parenchyma. The majority of recurrences after a negative second-look procedure are in patients with poorly differentiated cancers.