|
Advertisement |
|
|
||
 |
|
|||||
|
|||||
|
 |
||||||
|
||||||
|
||||||
|
© 2003 American Society for Clinical Oncology
The Myth of Measurable Disease in Ovarian CancerThe Cleveland Clinic Taussig Cancer Center and the Department of Hematology/Medical Oncology, The Cleveland Clinic Foundation, Cleveland, OH YOU HAVE just finished reading an article describing the results of a single-institution trial of a new cytotoxic agent being evaluated as a second-line ovarian cancer treatment strategy. The investigators conclude that the drug produced a "14.5% response rate in patients with measurable disease." You now ask yourself the following question: Does this result, by itself, have any clinical relevance? The answer is almost certainly "No." Why? We now confront the myth of measurable disease in ovarian cancer. Several generations of oncologists, both medical and gynecologic, have been indoctrinated into accepting the importance of the gold standard of phase II (and many phase III) clinical trials—the objective response rate of measurable tumor masses. In the recent past, for a patient to be considered to have achieved a partial response, it generally would be necessary for there to be "a 50% reduction in the sum of the products of maximum and perpendicular diameters of measurable lesions" appreciated during physical examination or visualized in radiographic evaluation.1 This objective measure of response has now been refined and replaced by the Response Evaluation Criteria in Solid Tumors, which defines a partial response as "at least a 30% decrease in the sum of the longest diameter of target lesions, taking as reference the baseline sum longest diameter."2 How reliable are these findings in ovarian cancer trials? Are they reproducible? Can they be audited? Are they of any direct relevance to the patient you are considering for treatment? Let us first consider the issue of measurable disease as determined by physical examination. More than 20 years ago, Moertel and Hanley3 asked 16 experienced oncologists to measure a number of simulated tumor masses. The physician subjects did not know that two of the masses were of identical size. As a result, the investigators were able to examine both intraobserver and interobserver variability in the measurement of this tumor. If we use the classic 50% definition of response noted above, there would have been an 8% or 7% objective response rate reported by the same investigator or different investigators, respectively, due solely to measuring error. If a 25% definition of response had been used (which is not that different from the Response Evaluation Criteria in Solid Tumors), the intraobserver and interobserver response rates due solely to measuring error would have been 19% and 25%, respectively.3 Other investigators have confirmed these general findings.4 A more recent experience, which specifically examined interobserver variability in the measurement of ovarian masses, reached similar conclusions.5 A group of obstetricians (n = 34), gynecologic oncologists (n = 26), and residents (n = 12) was asked to estimate the size of a simulated primary ovarian cancer and several metastatic lesions in the operative setting. The actual primary ovarian mass measured 13 cm in diameter, whereas the estimates of its size offered by the participating physician subjects ranged from 4 to 20 cm. A para-aortic node, the actual diameter of which measured 2 cm, was stated to be from 0 to 5 cm in size; a 5-cm mass in the hemidiaphragm was estimated at 2 to 7 cm; and an 8.5-cm right pelvic sidewall lesion was estimated at 0 to 12 cm in diameter. Of note, there were no differences in accuracy documented among the three groups of observers. Furthermore, it must be remembered that actual data provided from a physical examination cannot be audited, except for the mere fact that the physician has written a particular tumor size in the medical record. This documentation within the medical record is a poor substitute for objective quality control of clinical research. Did the mass appreciated on vaginal examination really shrink from 3 x 2 cm to 2 x 1 cm (a partial response by > 50% decrease in the product of the perpendicular diameters), or did it shrink only to 2 x 2 cm (no response)?
It also should be noted that patients considered for entry into primary chemotherapy trials are frequently divided into two broad groups: those with optimal residual (eg, largest tumor mass Similarly, independent audits of reported objective responses that were based on radiographic evaluation (eg, computed tomography scans or ultrasound) in ovarian cancer reveals that as many as 30% to 40% of responses are unable to be confirmed.6–8 For example, the response rate in a phase II trial of single agent, topotecan (N = 92 patients), decreased from 25.8% to 15.2% after independent radiographic review of the claimed responses.6 In a large phase II experience (N = 293 patients) that examined second-line, single-agent docetaxel in ovarian cancer, more than 10% of patients were found on external review to be "unevaluable because of equivocal radiological findings."7 Finally, in a randomized trial evaluating the optimal dose and schedule for paclitaxel in the second-line management of ovarian cancer (N = 382 assessable patients), independent external radiographic review eliminated eight of the 74 initially reported radiographic responses.8 Although it might be argued that the quality of all phase II to III ovarian cancer trials employing radiographic response as a study end point would be improved substantially through mandated independent review, the cost and effort associated with such an undertaking make it highly unlikely that this process will occur, except as required by regulatory bodies (eg, the United States Food and Drug Administration). Furthermore, even so-called experienced radiologists often disagree with measurements of tumor masses, particularly those with poorly defined and irregular lesions.9 Thus, if physical examination and radiographic evaluation are unable to provide a reliable and accurate assessment of the level of biologic activity of treatment programs being explored in patients with ovarian cancer, is there a superior alternative? Although currently it is not possible to provide a definitive answer to this question, it is relevant to note the potential superiority associated with the serum tumor maker, CA-125.10–12 Laboratory measurement of the CA-125 antigen level is relatively simple and inexpensive (as compared with a computed tomography scan), and highly standardized external quality control is feasible (including, if necessary, sending all samples to a single reference laboratory). Furthermore, the test provides a simple number for purposes of comparison (rather than complex measurements of diameter or length and width), it is easy to audit, and unexpected results (eg, major declines or elevations that may represent a laboratory error) can be rapidly repeated. Finally, this strategy eliminates observer bias and inexperience, as well as the previously noted substantial variations caused by both intraobserver and interobserver interpretation of physical findings and radiographic evaluation. The use of CA-125 to evaluate tumor response has been criticized because changes in the antigen level may be discordant with increases or decreases in the size of measurable disease.13,14 However, an alternative explanation for such findings may be that the CA-125 antigen level provides a more global assessment of the extent of tumor volume, in contrast to the limited view of total body cancer burden gleaned from observation of changes in size of a localized malignant mass lesion.15 Is it really rational to think that what happens to a 2-cm nodule palpated on pelvic examination will be more reflective of the overall status of the cancer in an individual patient than major changes in the serum level of CA-125? In addition, discordance between the size of a growing individual malignant lesion and a declining CA-125 antigen level can appropriately be viewed as an example of the increasingly recognized heterogeneity of tumor within a single patient.15 Why should it come as a surprise to anyone that the overall tumor burden might be decreasing in response to chemotherapy (resulting in a decline in the CA-125 antigen level), while an individual mass lesion containing a preponderance of cells with a chemotherapy-resistant phenotype may continue to grow? Finally, as with all blood markers, it is important to recognize the limitations of the CA-125 antigen level in monitoring the course of disease, including the possibility of laboratory error, the influence of intercurrent medical illness (eg, peritonitis), and the effects of surgery (eg, tumor cytoreduction that results in a decrease in the tumor marker and peritoneal irritation, which in turn causes an increase in the antigen level). It can realistically be anticipated that during the next decade there will be a substantial increase in the number of biologically active antineoplastic agents used as second-line therapy in ovarian cancer, with reported response rates of measurable tumor masses in excess of 10% to 15%. In this setting, and in the absence of data from definitive randomized phase III trials, it will be tempting for oncologists, and patients, to succumb to the myth of measurable disease, and select treatment that is based solely on the drug that is stated to possess the greatest opportunity to achieve an objective tumor response. Although evidence of biologic activity is certainly of interest, in the absence of data from randomized trials, it will be unknown if those individuals achieving a partial response experience any benefit compared with patients considered to have stable disease. It is important that all oncologists recognize the limitations inherent in our current ability to use this information to develop a rational management strategy. Finally, the relevance of any response data, whether reported from phase II or III trials, and the use of criteria of measurable tumor masses or favorable changes in a surrogate marker will be considerably strengthened if such activity is clearly shown to correlate with a decrease or elimination of cancer-related symptoms and an improvement in the patient’s quality of life. It is an error simply to assume that documented evidence of biologic activity of an antineoplastic agent automatically translates into clinical benefit for a patient. REFERENCES 1. Miller AB, Hoogstraten B, Staquet M, et al: Reporting results of cancer treatment. Cancer 47:207–214, 1981[CrossRef][Medline]
2. Therasse P, Arbuck SG, Eisenhauer EA, et al: New guidelines to evaluate the response to treatment in solid tumors. J Natl Cancer Inst 92:205–216, 2000 3. Moertel CG, Hanley JA: The effect of measuring error on the results of therapeutic trials in advanced cancer. Cancer 38:388–394, 1976[CrossRef][Medline] 4. Warr D, McKinney S, Tannock I: Influence of measurement error on assessment of response to anticancer chemotherapy: Proposal for new criteria of tumor response. J Clin Oncol 2:1040–1046, 1984[Abstract] 5. Prefontaine M, Gelfand AT, Donovan JT, et al: Reproducibility of tumor measurement in ovarian cancer: A study of interobserver variability. Gynecol Oncol 55:87–90, 1994[CrossRef][Medline]
6. Gwyther S, Bolis G, Gore M, et al: Experience with independent radiological review during a topotecan trial in ovarian cancer. Ann Oncol 8:463–468, 1997 7. Kaye SB, Piccart M, Aapro M, et al: Docetaxel in advanced ovarian cancer: Preliminary results from three phase II trials. Eur J Cancer 31A:S14–S17, 1995
8. Eisenhauer EA, ten Bokkel Huinink WW, Swenerton KD, et al: European-Canadian randomized trial of paclitaxel in relapsed ovarian cancer: High-dose versus low-dose and long versus short infusion. J Clin Oncol 12:2654–2666, 1994
9. Hopper KD, Kasales C, Van Slyke MA, et al: Analysis of interobserver and intraobserver variability in CT tumor measurements. AJR Am J Roentgenol 167:851–854, 1996
10. Rustin GJS, Nelstrop AE, McClean P, et al: Defining response of ovarian carcinoma to initial chemotherapy according to serum CA 125. J Clin Oncol 14:1545–1551, 1996
11. Bridgewater JA, Nelstrop AE, Rustin GJS, et al: Comparison of standard and CA-125 response criteria in patients with epithelial ovarian cancer treated with platinum or paclitaxel. J Clin Oncol 17:501–508, 1999
12. Rustin GJS, Nelstrop AE, Bentzen SM: Selection of active drugs for ovarian cancer based on CA-125 and standard response rates in phase II trials. J Clin Oncol 18:1733–1739, 2000 13. Davelaar EM, Bonfrer JMG, Verstraeten RA, et al: CA 125, a valid marker in ovarian carcinoma patients treated with paclitaxel? Cancer 78:118–127, 1996[CrossRef][Medline] 14. Morgan RJ Jr, Speyer J, Doroshow JH, et al: Modulation of 5-fluorouracil with high-dose leucovorin calcium: Activity in ovarian cancer and correlation with CA-125 levels. Gynecol Oncol 58:79–85, 1995[CrossRef][Medline] 15. Markman M, Kennedy A, Webster K, et al: Declining CA-125 in an ovarian cancer patient with progression of measurable disease: A rational hypothesis for discordant results. Gynecol Oncol 77:321–322, 2000[CrossRef][Medline]
This article has been cited by other articles:
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|||||||||||
|
|||||||||||
|
|
Copyright © 2003 by the American Society of Clinical Oncology, Online ISSN: 1527-7755. Print ISSN: 0732-183X
|
1 cm in maximal diameter) or those with suboptimal residual (eg, largest tumor mass > 1 cm) disease. Although the overall extent of residual disease clearly has been shown to influence prognosis in ovarian cancer, how are the results of trials potentially influenced by the inadequacy of tumor measurements? 
