Emerging Chemotherapy Response Predictors

The clinical importance of predicting who will and who will not respond to chemotherapy is intuitively obvious. If a test could predict who will respond to a given drug, the treatment could be administered only to patients who benefit, and others could avoid the unnecessary treatment and its toxicity. However, the practical development of chemotherapy response prediction tests poses several challenges. There are theoretical limits to the accuracy of any response predictor which measures the characteristics of the cancer only. Host characteristics which are not easily measured in cancer tissue, including the rate of drug metabolism, can have an important impact on response to therapy. Also, there is considerable uncertainty as to what level of predictive accuracy would be clinically useful. In fact, different levels of predictive accuracy may be required for different clinical situations.

For instance, the clinical utility of a chemotherapy response prediction test that has 60% positive predictive value (PPV; i.e. 60% chance of response if the test is positive) and 80% negative predictive value (NPV; i.e. 20% chance of response if the test is negative) will depend not only on these test characteristics but also on the availability and efficacy of alternative treatment options, as well as the frequency and severity of adverse effects, and the risks of exposure to ineffective therapy (i.e. rapid disease progression with life-threatening complications). A test with the above performance characteristics may be of limited value in the palliative setting, when alternative treatment options are limited and generally ineffective. Patients and physicians may want to try a drug even if the expected response rate is only 10% (well within the range of test negative cases), particularly if side-effects are uncommon or tolerable. On the other hand, in the setting of potentially curative therapy, when multiple treatment options are available, a test with the same performance characteristics may be helpful to select the best regimen from the several treatment options. In addition, a test developed to predict response to a given treatment in previously untreated patients may not predict response sufficiently accurately when the same drug is used as second- or third-line treatment.

Considering these complexities, not surprisingly many of the recent predictive marker studies that employed high-throughput analytical tools focused on the preoperative (neoad-juvant) treatment setting in breast cancer. Neoadjuvant chemotherapy provides a unique opportunity to identify molecular predictors of response to therapy. Pathologic complete response (pCR) to chemotherapy indicates an extremely chemotherapy-sensitive disease and represents an early surrogate of long-term benefit from therapy. Histological type, tumor size, nuclear grade and ER status all influence the probability of response to neoadjuvant chemotherapy, and these clinical variables can be combined into a multivariable model to predict probability of pCR (http://www. mdanderson.org/care_centers/breastcenter/

dIndex.cfm?pn=448442B2-3EA5-4BAC983100 76A9553E63).15 However, these clinical variables lack regimen-specific predictive value and represent features of general chemotherapy sensitivity.

Several small studies provided "proof-of-principle" that the gene expression profile of cancers which are highly sensitive to chemotherapy are different from tumors which are resistant to treatment.16 The largest study so far included 133 patients with stage I—III breast cancer who all received preoperative weekly paclitaxel and 5-fluorouracil, doxorubicin, cyclophosphamide (T/FAC) chemotherapy.17 The first 82 cases were used to develop a multigene signature predictive of pCR and the remaining 52 cases were used to test the accuracy of the predictor. The overall pCR rate was 26% in both cohorts. A 30-gene predictor correctly identified all but one of the patients who achieved pCR (12 of 13) and all but one of those who had residual cancer (27 of 28) in the validation set. It showed significantly higher sensitivity (92% vs 61%) than a clinical variable-based predictor including age, grade and ER status. The high sensitivity indicates that the predictor correctly identified almost all of the patients (92%) who actually achieved pCR. The PPV of the pharmacogenomic predictor was 52% (95% CI 30-73%); however, the lower bound of the 95% CI did not overlap with the 26% pCR rate observed with this regimen in unselected patients. This indicates that the predictor could define a patient population that is more likely to achieve pCR than unselected patients. The NPV of the test was also high, 96% (95% CI 82-100%), which indicates that <5% of test-negative patients (i.e. predicted to have residual disease) achieved pCR. These performance statistics are similar, with regard to the NPV and better with regard to PPV, than those seen with ER immunohistochemistry or HER2 gene amplification as predictive markers to endocrine or trastuzumab therapies, respectively. However, to what extent this genomic predictor of sensitivity is specific to T/FAC therapy rather than being a generic marker of chemotherapy sensitivity is yet to be determined.

10 Ways To Fight Off Cancer

10 Ways To Fight Off Cancer

Learning About 10 Ways Fight Off Cancer Can Have Amazing Benefits For Your Life The Best Tips On How To Keep This Killer At Bay Discovering that you or a loved one has cancer can be utterly terrifying. All the same, once you comprehend the causes of cancer and learn how to reverse those causes, you or your loved one may have more than a fighting chance of beating out cancer.

Get My Free Ebook


Post a comment