I’m currently attending the 2009 American Society of Clinical Oncology (ASCO) breast cancer symposium, in San Francisco.
The Keynote Address (by Martine PIccart-Gebhart of the Jules Bordet Institute) was very relevant to Individualized Tumor Response Testing (ITRT). The speaker made the point that only 8% of new drugs entering Phase I trials ever make it to marketing and that this percentage is even lower for cancer drugs, because “current drug testing is inefficient, with many drugs failing late in development,” with “these expensive failures” owing, in large measure, to “ineffective drugs” and “poor patient selection (i.e, lack of prognostic and predictive markers for response to therapy).”
The speaker went on to note that “little progress has been made in identifying which therapeutic strategies are likely to be effective for individual patients.” The speaker concluded that “identifying markers that can predict response to a particular drug remains a great challenge.”
To my knowledge, there is absolutely nothing being presented at this meeting which reports any progress at all in drug selection through the use of molecular profiling. Perhaps I’ve missed something. There is another half day remaining.
When microarrays and high throughput RT-PCR emerged some years back, I’d have predicted quite a bit of progress by now. Instead, there were a couple of early clinical studies, and virtually no follow up by now. Certainly, the sum total of the cell culture-based literature is vastly superior, taken as a whole, to the sum total of the molecular profiling-based literature for drug selection. One would think that there would be at least equal interest in cell culture-based approaches. Sadly, no.
I presented what I think was an important study in a poster session. It attracted virtually no interest. I’d like to briefly go through the most important points here. I’d be very interested in feedback, comments, and/or criticism, preferably on this blog, but, if preferred, as a private email: email@example.com
The problem I addressed is that of so-called “triple negative” breast cancer (TNBC), which is: Estrogen Receptor Negative (ER-), Progesterone Receptor Negative (PR-), and Her2 Negative (H2-).
When breast cancer presents as locally advanced disease, it is customarily treated with neoadjuvant (preoperative) chemotherapy, followed by definitive surgery. At the time of surgery, the specimen is assessed to determine if all visible tumor has been destroyed by chemotherapy. When this happens, it is said to be a “pathological complete response” (pCR). The Kaplan-Meier curves below show survival in Non-TNBC and TNBC, as a function of whether or not preoperative chemotherapy produced a pCR.
In the figure below, the yellow line is the survival for non-TNBC patients who achieved a pCR; the blue dashed line is the survival of TNBC patients who achieved a pCR; the black dashed line is for non-TNBC who did not achieve a pCR; and the red dashed line is for TNBC patients who did not achieve a pCR.
These data show that obtaining a pCR is everything. Get a pCR, and the survival is excellent. Don’t get a pCR and the survival, especially for TNBC patients, is very poor.
So the issue is this: what can we do to improve the pCR rate in TNBC?
I began by going through our database of breast cancer cell culture assays (using cell death endpoints) and trying to see if there were any drugs which appeared to be uniquely active in TNBC vs non-TNBC. Here’s a representation of what I found:
The horizontal bars show percent control cell survival for drug exposed cultures. The shorter the bar, the more active the drug. Error bars represent 95% confidence limits. These are representative data for a few of the more important drugs.
Essentially, there are no major differences between the activity of most drugs in TNBC vs non-TNBC, with one glaring exception: cisplatin, which was dramatically more active in TNBC than in non-TNBC.
I went on to present a lot of data further dissecting which specific markers were most associated with TNBC sensitivity to cisplatin. These data will shortly appear on the ASCO website, when they post slide presentations from the 2009 Breast Cancer Symposium. But the following were the most important “take home” messages:
Major markers for platinum sensitivity in breast cancer were:
1. Estrogen Receptor negativity
2. Bloom-Richardson (Nottingham) score of 9/9 (i.e. very poorly differentiated tumors.
Minor markers for platinum sensitivity in breast cancer were:
1. Her2 negativity
2. Progesterone receptor negativity
The following puts the findings into a useful perspective:
These data compare TNBC to other types of tumors. It’s known that renal cell carcinomas are very resistant to cisplatin (less than 10% response rate), and that is reflected by the cell culture (cell death endpoint) data. It’s known that previously-untreated, poorly differentiated ovarian cancers tend to be very sensitive to cisplatin (70% response rate), and that’s also reflected by the cell culture data. When ovarian cancer patients relapse soon (0 to 6 months) after discontinuation of chemotherapy, they have only a 25% response rate to re-treatment with platinum. When ovarian cancer patients relapse greater than 6 months following discontinuation of chemotherapy, they have a 50% response rate to re-treatment with platinum. These clinical findings are also nicely recapitulated by the cell culture assay data.
Now, breast cancers which are either estrogen receptor positive and/or more than very poortly differentiated (Bloom Richardson score of 4 to eight) tend to be even more resistant to cisplatin than are previously treated ovarian cancer which relapse soon (0 to 6 months) after discontinuation of chemotherapy. In contrast, Triple Negative Breast Cancers tend to be as sensitive or more sensitive (especially when also Bloom Richardson 9/9) to cisplatin than are previously-untreated, poorly-differentiated ovarian cancers.
I think that these data clearly show the utility of cell culture assays in “targeting” chemotherapy to patient sub-groups who are most likely to benefit from treatment with given individual drugs. It’s hard to see how “molecular” profiling tests could have produced similar insights.
The following were my summary conclusions, comparing cell culture profiling versus “molecular” profiling:
So the question is this: Why is there absolutely no interest in cell culture tests?
- Larry Weisenthal/Huntington Beach, CA