What about carboplatin? There have been three trials internationally looking at carboplatin. A British study comparing AUC6 to AUC12. A Danish study comparing AUC4 to AUC8 and an Austrian study comparing cisplatin to a combination of cisplatin plus carboplatin. Again, no advantages to increasing the dose intensity of the platinum compound by escalation of carboplatin dose or by its addition to cisplatin. So it would appear, based on this, there is no reason to go above an AUC of 4 with carboplatin.

What about Taxol? Well Bristol-Meyers Squibb funded Johns Hopkins to do a metaanalysis of the first five phase II trials of Taxol in ovarian carcinoma. That was presented in 2006 at ASCO by Eric Lewinsky. What that showed was that there was an improved response rate with an increasing dose of Taxol up to a dose of 175 mg per meter square every three weeks. Beyond that there was actually a decrement in response for each additional 10 mg per meter square increase. We’ve had two randomized trials looking at this issue, one comparing Taxol 135 per meter square to Taxol 175 per meter square; 15% response rate with 135, 20% response rate with 175, not significantly different. Then another study comparing 175 to 250, a GOG trial, with response rate going from 27.5% at 175 to 36% at 250. No difference in progression-free survival and no difference in overall survival. If there is a dose-response relationship with Taxol, the slope of that curve is not steep at all but the slope of the toxicity curve is dramatic. So the best evidence we have today would appear to suggest that there is no reason to go beyond 175 mg per meter square of Taxol. So these would represent reasonable dose levels for these three drugs. You may want to ask, “Well, gee, why did you put 60-70 per meter square and why did you put AUC 4-5 when you’ve shown evidence for 60 and 4?” Well, it’s that little bit in me that got trained that said more is better. I just can’t let go of it. And neither can anybody else. So the data suggests that we certainly don’t need to go to higher levels than these because all we do with that is increase toxicity.

What about Taxol schedule? We really don’t have hard data on a lot of this yet but I think that what we do know about Taxol schedule is the longer the infusion the greater the myelosuppression. The shorter the infusion the greater the non-hematologic toxicity. In fact with the longer infusions you don’t see cumulative toxicity with Taxol but with the shorter infusions you do see cumulative non-hematologic toxicity. The optimal schedule is not known. Either a three or 24-hour infusion is preferred right now because we have a lot of data on those. As of yet we don’t have any conclusive proof that a weekly one hour schedule is acceptable, although I know that a lot of you probably use that. And as of yet we don’t have much in the way of data in a 96-hour infusion, which is actually being tested in a randomized phase III trial now. Because, at least in breast cancer, there are some reasons to believe that a more prolonged infusion might be better and the preclinical data supports that. But right now either the three or 24-hour infusion, either one, would be a reasonable approach to the use of Taxol in ovarian carcinoma.

At that same meeting at ASCO the fly in the ointment was presented. That was GOG protocol 132. This study was done while we were waiting for the data on GOG 111 to mature. This randomized patients with large volume advanced disease to either cisplatin, Taxol or a combination of Taxol plus cisplatin, with the Taxol given as a 24 hour infusion. Again, each regimen was given every three weeks for a total of six cycles.

While we thought we knew what the results of this study would be before we started, and what we found out was that there is no significant different difference in overall survival. The Taxol single agent arm gave a lower response rate and a shorter progression-free survival, but overall survival was exactly the same. You might ask yourself, “Why did this study differ from OV10 and GOG111?” Well the circumstances were quite different. When GOG111 and OV10 were done Taxol was not available for salvage therapy in those countries. In the U.S. for 111 and in Canada and Europe for OV10. So patients who were assigned a Cytoxan/cisplatin did not get salvage Taxol. On the other hand, when this study was done, Taxol was commercially available and in fact what happened was that at the end of six cycles of therapy if patients had any residual disease left, they immediately got salvage therapy.

Patients assigned to the cisplatin arm got Taxol and patients assigned to the Taxol arm got cisplatin.

Patients assigned to the combination arm got a whole hodgepodge of different things, most of which probably did absolutely nothing or even harmed the patient.

So as a result, half the patients got some form of salvage therapy and we believe the crossover to the opposite drug accounts for the blunting of the survival differences that you see in the study. Also this was not merely salvage therapy. This was immediate use of the other drug without waiting for disease progression to take place. So in effect what we were really studying here was sequential cisplatin for six cycles followed by Taxol for 6-12 cycles or Taxol for six cycles followed by cisplatin for six cycles, or concurrent therapy. We proved something here that I think is going to turn out to be very useful for us. We’ve proved that you can give drugs concurrently or sequentially and achieve the same results. That sort of approach is going to be applied in trying to introduce additional active new agents to up front therapy, that is, sequential doublets will probably be used in that experimental approach.

Now the most recent study to be completed was ICON-3, the International Collaborators on Ovarian Neoplasms study number 3. This study was conducted in the United Kingdom and Italy. About 70% of the patients came from the United Kingdom, about 30% from Italy. Patients here were randomized to the experimental arm of Taxol/carboplatin, Taxol given as a three hour infusion followed by carboplatin. Or to one of two control arms. Each institution had to select which control arm the institution would use; either carboplatin alone or a three drug combination of Cytoxan, Adriamycin and cisplatin.

All stages of the disease were allowed onto this study. So this study includes stage I, stage II, stage III and stage IV patients. One-third of the patients in the study on the control arm also received salvage Taxol. So there will be some blunting of any survival differences as you would expect, because of the salvage therapy. Seventy percent of the control patients received single agent carboplatin as a control, 30% received a three drug combination. Now I should tell you that there was an ICON study, ICON-2, that showed no difference between the three drug combination and carboplatin so it’s reasonable to lump the control arms. Now if you attended ASCO and attended the GYN session you know that the ICON investigators presented the following data; in terms of progression-free survival, no difference between the two arms.

In terms of overall survival, no difference between the two arms. What they did not present was presented by Dr. McGuire in his discussion and was obtained from one of the ICON collaborators and that is a more detailed breakdown of the data. And this is a very important subset analysis; 622 of the patients had stage I disease. Only 83 events had occurred in that subset. Too few events to allow any meaningful conclusions to be drawn. But 1,452 of the patients had either stage III or stage IV disease and among those 1,452 there had been 544 events, enough events to allow valid conclusions to be drawn. In this subset, which is the same subset studied in OV10 and a similar subset to GOG111, hazard ratio was 0.82 and that is statistically significant at a P value of 0.03. So in the same population of patients studies in OV10 and GOG111 you see the same result, a benefit from concurrent Taxol, platinum administration. If you put all four of these trials on the same slide you will note that three of the four, GOG111, OV10 and ICON-3, show a clear value for the Taxol/platinum combination. One, GOG132, shows no difference but is accounted for the immediate rather than the salvage use of the opposite drug. The weight of evidence in these four studies speaks to the standard of care as being a combination of Taxol and platinum. And if you are asked that on the Boards, that’s what your answer is.

Now after you have completed the exploratory laparotomy you can stage the patient in more detail. This is the more detailed FIGO staging system for stage III and stage IV disease. What we’ve added here is a subdivision of stage III disease according to volume of residual disease; IIIa, those with microscopic disease only outside the pelvis, IIIb, macroscopic disease outside the pelvis but no nodule bigger than 2 cm in diameter, IIIc, nodules bigger than 2 cm in diameter outside the pelvis, or involvement of the retroperitoneal or inguinal nodes. The staging system refers to disease as it appears when the belly is open before the surgeon operates. The way we actually look at this disease clinically is to divide these patients into two groups; those who finish surgery with small volume residual disease – no nodule bigger than 2 cm remaining – and those who finish surgery with bulkier disease. According to the SERE in the United States, roughly 40% of all advanced disease patients will have small volume residual disease, about 60% will have large volume residual disease. But if you look just at the patients included in GOG studies, roughly 70% of the patients in surgery with small volume residual disease, 30% with bulky disease. I think it indicates the difference in having a gynecologic oncologist perform the surgery.
As of 1990 the standard of care for patients with bulky disease was a maximum attempt at cyto-reduction followed by chemotherapy. The standards at that time were Cytoxan/cisplatin or Cytoxan/carboplatin. What we want to look at is what has been done in the last nine years, in particular the last one to two years. There have been three themes that have dominated clinical research during the last decade. The integration of Taxol in to front line therapy, the choice of a platinum compound and the optimum dose of schedule of the drugs that need to be used. With regard to the first of these, the introduction of Taxol into front line therapy; there have now been four major randomized trials completed and reported. The most recent in May of this year at ASCO. GOG protocol 111, EORTC NCIC OV10, GOG protocol 132 and ICON-3. We want to look very briefly at each of these four.

GOG protocol 111 was the first of the studies to be published. This was published in the January 4, 2006 issue of the New England Journal of Medicine. This study took patients with large volume advanced disease, randomized them to receive either Cytoxan/cisplatin or Taxol as a 24 hour infusion followed by cisplatin with each regimen being given every three weeks for a total of six cycles of therapy. If you read the publication you know that in regard to the five major parameters looked at, Taxol/cisplatin was statistically superior to Cytoxan/cisplatin. Overall response rates, 73% versus 60%. Clinical complete response rate, 51% versus 31%. Percentage of patients grossly disease free at second-look laparotomy, 40% versus 24%. Median progression-free survival, 18 versus 13 months. Median overall survival 38 months versus 24 months. We now have follow-up out to nine-plus years and those differences continue to be maintained.

In 2006 at ASCO the results of the confirmatory trial was presented, this was OV10 a European/Canadian study. There are several differences in this study compared to the GOG study. Number one, patients with both large and small volume disease were included. Number two, the Taxol was given as a three hour infusion. Number three, up to nine cycles of treatment were allowed, and most patients did receive more than six. As an aside I should add that what the study also demonstrated was that you do not want to use three hour Taxol with cisplatin because there is an unacceptably high rate of significant neurotoxicity, either grade III or grade IV. That’s been demonstrated not only in this study but in two other trials that used three hour Taxol with cisplatin. What this study did was essentially reproduce the GOG study. In terms of overall response rate, Taxol/cisplatin was superior; 77% versus 66%. Clinical complete response rate, 50% versus 36%. Median progression-free survival, 16.6 versus 12 months, and median overall survival 35 versus 25 months.

Important prognostic factors in ovarian carcinoma include age, and one thing to note about age; unlike most other solid tumors, older patients with ovarian carcinoma develop more aggressive disease. The older the patient the more aggressive the tumor is likely to be. The younger the patient, the less aggressive the tumor is likely to be. The reference is a supplement to Cancer in 1992 which summarizes cooperative group data from all over the world. Stage is also an important prognostic factor and perhaps the most important prognostic factor. In fact, this will be the factor that we use to make treatment recommendations. Volume of residual disease in patients with stage III ovarian carcinoma is an important prognostic factor. In early stage disease histologic grade and, as we’ve already discussed, occasionally histologic type is an important prognostic factor. There are beginning to be some molecular biological factors identified as predictive of a poor outcome. Most important prognostic factor though is the FIGO stage, which is shown in abbreviated form here on this slide. Stage I disease is disease confined to the ovaries. This accounts for about 20% of ovarian carcinoma. Stage II disease is disease outside the ovaries but confined to the pelvis and accounts for about 5% of ovarian carcinoma. So roughly 1:4 patients will have what we would call limited or early stage disease. The most common stage at presentation is stage III disease which is spread by direct seeding throughout the peritoneal cavity, or involvement of the retroperitoneal or inguinal lymph nodes. Fifty-eight percent of all patients will have stage III disease at diagnosis. Then finally, stage IV disease which has spread to more distant sites accounts for about 17% of patients. Stage III and stage IV together account for what we call advanced ovarian carcinoma.

Now because this is an intraabdominal disease process in the vast majority of patients, as we’ve indicated, all but 17% of the patients will have disease confined to the peritoneal cavity at the time of diagnosis. Because of that, exploratory laparotomy is an extremely important step in the diagnostic evaluation of the patient and it is also the first step in treatment of the patient. Now I’m not a surgeon, I’m a medical oncologist so I’m not going to try to tell you how to operate on ovarian cancer patients, but everybody dealing with ovarian cancer has agreed on three basic principles for exploratory laparotomy. Number one; the incision should permit the exploration of the entire peritoneal surface. Number two; if there is no gross disease outside the pelvis, multiple biopsies should be taken. By the way, I should emphasize, this slide used to say “multiple blind biopsies”.

One of my surgical colleagues informed me that only interns do blind biopsies. Surgeons always see what they are biopsying so we took the word “blind” out in deference to the gynecologic oncologists. Finally, the surgeon should be prepared to undertake an aggressive attempt at surgical cyto-reduction. Now the basis for this last recommendation is mostly retrospective data in the literature suggesting that patients who start chemotherapy with small volume residual disease have better survival and better response to chemotherapy than patients with large volume residual disease. Until recently we had no proof from a randomized trial of this principle. However we do now. In 1995 the European Organization for Research in the Treatment of Cancer, EORTC, published the results of this study in which they took patients who had been initially operated on and deemed to be not bulk-reducible. They gave them three cycles of chemotherapy. At that time cisplatin and Cytoxan was the treatment of choice. They then randomized them to undergo interval surgical cyto-reduction or no interval surgical cyto-reduction, and then everybody got three additional cycles of chemotherapy. What that publication shows is that those patients assigned to interval debulking surgery – and all were included regardless of whether interval debulking surgery could be accomplished – of those patients assigned to interval debulking surgery there was a statistically significant improvement in progression free survival; 15 months versus 12.5 months, and an overall improvement in survival of 27 months versus 19 months.

The reason you see a discrepancy here between 408 and a total of 299 is that 109 patients had progressed before the end of three cycles of chemotherapy and were removed from study. Everybody who reached randomization point, those who got interval debulking surgery did better. We are in the process of confirming these data now in the United States with this study that will finish up in about a year, that uses Taxol and cisplatin as the treatment but otherwise follows the same design as the EORTC trial. That is, Taxol, cisplatin for three cycles then interval debulking surgery, then three more cycles of chemotherapy; versus simply six cycles of chemotherapy. As of right now, the weight of evidence would suggest that surgical cyto-reduction ought to be done – that’s the weight of both retrospective and prospective evidence – if you are going to do it, the rationale for doing this favors doing the debulking surgery up front rather than in the middle of chemotherapy or after chemotherapy has been completed.

The rationale is to remove resistant clones of cells and to remove large bulky tumor that is poorly vascularized. That is better done before you give your chemotherapy. So right now, the standard of care is up front surgical bulk reduction – and this last statement, as a medical oncologist, I can say this – surgery should be performed by a gynecologic oncologist because they are better at removing bulky tumor.

With regard to prophylactic oophorectomy; there have been publications in the literature, most notably from Buffalo, suggesting that prophylactic oophorectomy in women with positive family histories was indicated. But if you look at their own data published in 1993 in Cancer, what you find is that after prophylactic oophorectomy – and they had a total of 324 women who underwent prophylactic oophorectomy because of positive family histories – their incidence of primary peritoneal neoplasia at a median follow-up of six years was already 1.8% which exceeds the risk in the general population for ovarian carcinoma, and begins to approach the risk in patients with positive family histories. So it would appear that what we are doing here is removing only a small portion of the tissue at risk. Most of the tissue at risk, that is the coelomic epithelium that lines the peritoneal cavity is left in place. The risk of other problems after oophorectomy will increase. So as of right now the only group for which prophylactic oophorectomy is recommended for serious consideration is in those who have hereditary syndromes. And the recommendation was to consider that beginning at age 35 and later. And that came from the Ovarian Cancer Consensus Conference held in August 1994 at the National Cancer Institute. But that was based, as the conference report said, on no data, merely speculation that this would work. So as of right now there is really no truly scientifically defined role for prophylactic oophorectomy. There is a consensus role for those with hereditary syndromes.

Screening for ovarian carcinoma; if you read luminary medical journals like Cosmopolitan and Redbook you will find that screening is something you really ought to be doing for your ovarian cancer patient. What those journals recommend is a combination of pelvic examination by the physician once a year, serial CA125’s annually, and annual transvaginal sonography. But the actual figures on the use of these show that, first of all, pelvic examination has no impact. We know that from numerous studies. When you combine transvaginal sonography and CA125 done in a serial fashion, the positive predictive value for finding ovarian carcinoma is 26.8%, which is really quite good. But, most of those cases are going to be stage III at diagnosis. The positive predictive value for stage I and stage II disease is 9.8% and by most standards a value of greater than 10% is required before you can label an approach as a valid screening test. Not only that, but the studies that have been done to date show that for every one laparotomy that you do for ovarian carcinoma, you are going to do somewhere between 14 and 39 for benign causes that did not require exploration. So there really is a downside for screening for ovarian carcinoma. So at least for right now screening is not recommended, there is no scientific basis for using the combination of transvaginal sonography and CA125 to screen for ovarian carcinoma. There is an ongoing randomized trial that we hope will at least in part answer this question.

Pathologically, coelomic epithelial carcinomas of the ovary fit into four broad categories; serous, endometrioid, mucinous and clear cell. You will note that the two most common types are serous and endometrioid and that’s fortunate, because the two poor-risk types are mucinous and clear cell carcinomas. These two tumors have a much poorer prognosis than the serous and the endometrial types. If you look just at patients with stage III and stage IV disease what you find is less than 10% of the patients have either mucinous or clear cell carcinomas. Almost 90% will have either serous or endometrioid tumors. We don’t know what to do differently for the mucinous and clear cell tumors, currently, to hopefully impact greater on their outcome. So right now that histology has very little impact on how you are going to manage the patient clinically, except to know that the patient is at much greater risk for death due to disease if they have a mucinous or clear cell carcinoma. The only part of histology that really makes a difference in our approach to the tumors is if the pathologist identifies a particular ovarian cancer as being a borderline ovarian carcinoma. That is, it appears to be a well-differentiated tumor with no underlying stromal invasion. These tumors have an excellent prognosis, even if you simply observe them, do nothing to treat them other than an attempt at surgical resection. And even those with intraabdominal spread, that is stage III disease, have an 80% survival at five years. In a large study done by the GOG, no stage I ovarian borderline tumor has occurred as yet at ten years of follow-up. So the treatment for borderline tumors is resection if possible, and then no further treatment until the tumor declares itself as a more aggressive tumor. At that point you would introduce systemic therapy.

In terms of molecular biology, we don’t have a lot of hard information on ovarian carcinoma. Things that you probably ought to take note of are that any mutations of P53 seen in ovarian carcinomas are probably a late change and have nothing to do with etiology. Number two, ras mutations are associated with borderline tumors of the ovary, which we will comment on in just a few moments. And number three, the frequency with which HER2/neu is over expressed is probably considerably lower in ovarian carcinoma than is the case in breast cancer. The GOG currently is running a trial looking at Herceptin ovarian carcinoma. We’ve had to screen 550 patients just to find 25 who are either 2+ or 3+ or HER2 and therefore candidates for clinical trial. And out of those 25 there has been only one responder to Herceptin. So it would appear that HER2/neu is going to be much less of an important factor in ovarian carcinoma than appears to be the case in breast cancer. As yet we have no unifying model relative to molecular factors.

In terms of environmental factors; there are a number of proposed risk factors, none of which have been confirmed. High dietary intake of meat and animal fat has been associated. A high ratio of lactose consumption to red blood cell galactose 1-phosphate uridyltransferase has been postulated. Again not proven. Consumption of coffee, tobacco and alcohol have all been studied. There are some positive reports in the literature but there are a lot of others that deny an association. Then finally, there have been associations reported with industrial exposure, radiation and talcum powder used at surgery. Also mumps. But none of these are confirmed. There are no clear environmental factors that have been directly related to the disease process as yet.

In the area of prevention, there have been four methods proposed; use of oral contraceptives, the use of fenretinide, tubal ligation and prophylactic oophorectomy. And we’ll dispense with these middle two very quickly. There is an ongoing trial in Europe looking at fenretinide and there will soon be a randomized trial in the United States looking at this as a means of preventing ovarian carcinoma, but as yet we have no proof of efficacy. Tubal ligation is associated with a clear reduction in the risk of ovarian carcinoma, and this gets back to what we talked about earlier. That it is reducing the exposure of the peritoneal cavity to external agents. At least we think that’s the case. There is a clear reduction in risk with tubal ligation.

The other two topics have been looked at a good bit more in the literature. First of all, oral contraceptives; it’s clear that a five-year, five consecutive years of use, of oral contraceptives will reduce the risk of nulliparous women to that of multiparous women who don’t use birth control pills. So the use of birth control pills for five or more consecutive years is associated with a risk reduction. The use of birth control pills for ten or more years by patients with a positive family history will reduce their risk to below that of patients who don’t use oral contraceptives and who have negative family histories. The hazard ratio for women with positive family histories who use birth control pills as compared to those who do not is 0.5. A 50% reduction at 10 years, in the risk of developing ovarian carcinoma. Now there are no hard and fast recommendations on whether or not to use oral contraceptives. More importantly, we do not have good studies on the effects of oral contraceptives in those patients who have hereditary syndromes. So we do not know what the impact will be in that specific subset. So any decision regarding the use of oral contraceptives would have to be made in consultation with the patient and would also have to take into account other risks, such as the 1.46 hazard ratio associated with breast carcinoma in the ten or more years consecutive use of oral contraceptives.