|
Advertisement |
|
|
||
 |
|
|||||
|
|||||
|
 |
||||||
|
||||||
|
||||||
|
Journal of Clinical Oncology, Vol 25, No 24 (August 20), 2007: pp. 3615-3620 © 2007 American Society of Clinical Oncology. DOI: 10.1200/JCO.2006.09.4540 An Early Signal of CA-125 Progression for Ovarian Cancer Patients Receiving Maintenance Treatment After Complete Clinical Response to Primary Therapy
From the Southwest Oncology Group Statistical Center, Seattle WA; University of Arizona Cancer Center, Tucson, AZ; University of California, Irvine, Orange, CA; Gynecologic Oncology Group Statistical Center, Buffalo, NY; and the M.D. Anderson Cancer Center, Houston, TX Address reprint requests to Southwest Oncology Group (SWOG-9701), Operations Office, 14980 Omicron Dr, San Antonio, TX 78245-3217; e-mail: pubs{at}swog.org
Purpose For ovarian cancer patients receiving maintenance treatment after complete clinical response to primary therapy, we tested the predictive value of the following early signal of progressive disease (EPD) criterion based on serum CA-125: for patients with CA-125 nadir  10 U/mL, a confirmed value of  20 U/mL serves as an early signal of CA-125 progression; for patients with nadir more than 10 U/mL, a value 2x nadir that is confirmed predicts progression. Patients and Methods The EPD criterion was tested on Southwest Oncology Group trial 9701/Gynecologic Oncology Group trial 178 patients (n = 288) and compared with Gynecologic Cancer Intergroup criterion.
Results For 204 patients with known progressive disease, the progression date was predated by EPD by Conclusion Initial testing of the proposed CA-125 criterion resulted in a low false-positive rate and early prediction of disease progression in more than 50% of the patients tested. The proposed criterion may better reflect the timing of disease progression and should be investigated further.
For ovarian cancer patients undergoing primary treatment, the Gynecologic Cancer Intergroup (GCIG) CA-125 disease progression criterion1 has been widely used in practice. For patients with normal baseline CA-125, or whose values normalize during treatment, the criterion defines CA-125 2x upper normal limit (ULN) documented on two consecutive occasions as disease progression. For those patients whose CA-125 never normalizes, CA-125 2x nadir on two successive occasions defines disease progression. The most commonly used ULN is 35 U/mL.
Using this criterion along with Response Evaluation Criteria in Solid Tumors (RECIST), we previously confirmed the findings by Crawford et al2 that ovarian cancer patients with premaintenance therapy CA-125
Given that different prognostic groupings exist within the commonly regarded normal CA-125 range, the aim of the current exercise was to investigate, before meeting the GCIG progression criterion, whether there are early indications of disease activity that are dependent on the nadir CA-125 value achieved. In the setting of patients who achieve complete clinical response after primary treatment, we propose that the following CA-125 behavior may constitute an early signal of progressive disease (EPD): CA-125
Patient Population The details of progression-free survival results of SWOG 9701/GOG 178 were reported previously.6 Briefly, stage III or IV ovarian cancer patients who achieved a clinically defined complete response to primary therapy based on the combination of platinum and paclitaxel were randomly assigned to receive 12 or 3 cycles of monthly treatment of single-agent paclitaxel (175 mg/m2 during 3 hours). A premaintenance therapy baseline CA-125 level of 35 U/mL obtained within 28 days before study entry was required. The definition of disease progression included both the standard RECIST criterion and the GCIG CA-125 criterion (with 35 U/mL as the ULN), whichever occurred first.
Statistical Methodology
Characteristics of the patient sample included in this analysis, distribution of premaintenance therapy CA-125 levels, and existing disease progressions according to the GCIG CA-125 progression plus RECIST criteria are outlined in Tables 1, 2, and 3, respectively. According to the GCIG criterion, with ULN = 35 U/mL, 63% (129 of 204) of the disease progressions were first detected by RECIST only and 37% (75 of 204) were first detected by increased CA-125 values, either alone (25%) or jointly with RECIST (12%).
Given that CA-125 values meeting the GCIG criterion also satisfy the proposed EPD criterion, sensitivity of the latter is by definition 100% in the group of existing patients with progressive disease, and the timing of EPD can only be earlier. Table 4 summarizes the changes in disease progression date when applying the proposed early signal criterion.
Of the disease progressions detected by RECIST only, 39% could have been detected by the EPD criterion first. In other words, CA-125 values did start to increase in these patients but the increase failed to meet the GCIG criterion before disease progression by RECIST was noted. Notably, with the GCIG criterion, 63% (129 of 204) of the disease progressions were detected by RECIST only. However, with the proposed EPD criterion, 39% (79 of 204) of the disease progressions were detectable by RECIST only, without any indication from CA-125 values. In fact, had CA-125 been measured more regularly, this proportion is likely to have been lower. Of the 79 disease progressions detected by RECIST only, 49 had extended periods with no CA-125 data before progression, and marker values met the initial EPD criterion in 11 patients with no confirming measurements before disease progression. EPD concurred with RECIST or clearly did not occur before RECIST in only 19 of the 79 patients.
For disease progressions detected by the GCIG CA-125 criterion, the proposed early signal predated the progression in 85% of the patients: 47% by
For the 204 disease progressions overall, the EPD criterion gave an early signal in 56%: 30% by
However, the 10% (21 patients) in whom the EPD was more than 180 days earlier than the progression date bear close examination, given that the earlier signal potentially could be false. Of these, the time gaps between EPD and progression in 11 patients were related to long disease assessment intervals. Specifically, eight patients had a clear increase in CA-125, with slightly longer than 90-day assessment intervals, resulting in a gap in EPD to disease progression from 181 to approximately 200 days. In one other patient, CA-125 first met the EPD criterion when the value reached 34 U/mL (nadir = 13 U/mL); the next two assessments were 46 and 365 U/mL (disease progression) at 4 months and an additional 5 months later, respectively, with no interim values available. Another patient's CA-125 met the EPD criterion with two successive values of 22 and 26 U/mL (nadir = 8 U/mL), respectively; her only subsequent CA-125 assessment was 52 U/mL 6 months later, and disease progression according to RECIST was documented 3 months thereafter. Finally, one patient died (with disease progression by existing criterion) 9 months after EPD with no CA-125 values available for the 5 months before death, during which persistent disease was noted by pathology from a hernia repair. Disease progression could have been detected sooner had there been more frequent assessments for these 11 patients.
In addition, two patients had gradual but persistent increases in CA-125 values. Two other patients had CA-125 elevation above 70 U/mL within 3 months of EPD, whereupon one patient was treated with platinum plus gemcitabine and the other patient was treated with paclitaxel plus carboplatin. Technically, the CA-125 elevation status was not confirmed as required by the GCIG criterion, and the protocol-specified disease progression definition was not met until much later after these additional treatments in both patients. One other patient's CA-125 remained at approximately 10 U/mL (nadir = 9 U/mL) for 7 months after study registration. Her CA-125 then doubled and stayed mostly in the range of 21 to 26 U/mL (meeting the EPD criterion) with a single value of 55 U/mL for the next 8 months; during this time a possible lung lesion was noted and patient complained of severe fatigue. Her available CA-125 doubled again wtih values of 40, 41, The remaining five patients potentially could be construed as having false positive results. CA-125 decreased from 29 and 21 U/mL in one patient and 24 and 29 U/mL in another to less than 10 U/mL immediately after the EPD signal with no cancer treatments documented. However, one patient received corticosteroids and the other received nonsteroidal anti-inflammatory drugs. In both patients, subsequent reapplication of the EPD criterion would have predicted progressive disease unambiguously. The nadir of 11 U/mL for a third patient was an isolated value; the next lowest value was 15 U/mL. After the EPD criterion was satisfied, the patient's CA-125 value had a 9-month stable period in the range of 19 to 27 U/mL before increasing to successive values of 41, 64, 72, and 78 U/mL. Had the second lowest value of 15 U/mL been regarded as the nadir, the EPD criterion would not have been met until CA-125 reached 41 and 64 U/mL, at which point disease progression would be indicated properly. CA-125 for the fourth patient remained less than 10 U/mL throughout the 11-month treatment period, with a nadir of 6 U/mL. Once off protocol treatment, her CA-125 values were 21, 21 (meeting EPD criterion), 26, 27, 31, 53, and 191 U/mL (also meeting RECIST progression), with a 9-month period between the two values of 21 U/mL qualifying for EPD, and 3- to 4-month intervals between all subsequent measurements. The fifth patient's CA-125 increased from less than 10 U/mL to 21 to 23 U/mL, meeting the EPD criterion; her values then stayed between 20 to 36 U/mL for the next 11 months. CA-125 data were not available for the next 17 months before RECIST progression was recorded. Because of the retrospective nature of this analysis, no additional information or data are available; we conservatively consider disease progression predicted by the EPD to be false for these five patients. Among the 84 patients for whom disease progression was not documented, CA-125 data met the EPD criterion for nine patients. All available CA-125 values for these nine patients are shown in Figure 2. Data in Figures 2A, 2B, and 2C reasonably can be regarded as indicating patients who have subsequently developed progressive disease. The patient whose data are shown in Figure 2D was treated with tamoxifen on meeting the EPD criterion, and the patient whose data are shown in Figure 2E was lost to follow-up beyond the last known CA-125 date. True progressive disease status cannot be determined for these two patients. Disease progression was not reported for patients with data in Figures 2F, 2G, 2H, and 2I, with 10, 14, 30, and 39 additional months of follow-up after meeting the EPD criterion, respectively. Thus, these four patients have potential false-positive results.
Overall, 135 patients met the proposed CA-125 EPD criterion; 126 patients were among those with documented disease progression and nine patients were new occurrences. Nine results were potentially false positive (five existing disease progressions and four new calls by the EPD criterion) and two patients were indeterminate. The positive predictive value of the proposed EPD criterion was 124 of 133, or 93% (95% CI, 88% to 97%).
Appendix Table A1 (online only) lists the duration between EPD and disease progression by baseline CA-125 categories. As discussed, the group with no difference designated included many patients for whom the lack of CA-125 data could have lead to the lack of observed EPD or GCIG progression (or both). Therefore, rather than no difference, the true time between EPD and GCIG/RECIST progression is mostly unknown for this group. Excluding the patients with no difference designated, among those with EPD observed in this analysis, there was no indication that time from EPD to disease progression differed significantly by baseline CA-125 categories. The median time from EPD to progression was 62, 56, and 40 days for baseline CA-125
In a group of ovarian cancer patients in complete clinical remission, Santillan et al7 found "a progressive low-level increase" in serum CA-125 levels to be strongly predictive of radiographic disease recurrence. The specific patterns of increase investigated were an absolute increase of 5 U/mL from nadir, an absolute increase of 10 U/mL from nadir, and doubling of the CA-125 nadir. Our proposed EPD criterion is based on the GCIG CA-125 disease progression criterion, but the upper normal limit was lowered from 35 to 10 U/mL. Coincidentally, it satisfies the last two patterns of increase described by Santillan et al7 (an absolute increase of 10 U/mL from nadir, and doubling of the CA-125 nadir). Because of the retrospective nature of the current analysis, including sparse CA-125 data in many patients and a lack of explanations for unusual marker behavior in others, it was difficult to draw any firm conclusions regarding the true false-positive prediction rate of the proposed EPD criterion. We have chosen to take a conservative approach and, in situations in which it was less than certain that progression had occurred or that EPD was caused by noncancerous conditions, we have classified the early signal as possibly being a false-positive signal. Despite this fact, the positive predictive value of the proposed criterion was 93% (95% CI, 88% to 97%), a quite respectable preliminary finding considering the retrospective limitation of this evaluation.
The proposal for an early indication of CA-125 progression and the outcome of this analysis leads to some obvious questions. First, what is the need even to consider this early detection criterion? As previously noted, several retrospective examinations of patients with nadir CA-125 levels within the normal range have now rather clearly shown the important prognostic impact of this laboratory value even when the finding is considered to be in the normal range (generally Thus, from the perspective of conducting clinical trials, it would be inappropriate to consider that all patients whose nadir CA-125 level declines into the normal range have an equivalent risk of subsequent disease progression. Future trials should take this fact into consideration and stratify maintenance treatment candidates accordingly for trial design and analysis. In addition, it begs the question whether the same CA-125 progression criterion applies to patients with varying levels of marker normalization under primary therapy, or whether an additional nadir-dependent criterion may possibly better reflect disease activity.
Based on the observed predictive value of the EPD criterion and its timing, it is our preliminary assessment that its use will likely permit a more critical assessment of the biologic activity of a primary treatment program. Using the earlier occurrence of EPD or RECIST disease progression as the hypothetical onset date for progression, we analyzed progression-free survival with this end point (EPFS) for SWOG 9701/GOG 178. There were no qualitative differences between EPFS and PFS by GCIG/RECIST for either the 12- v 3-course paclitaxel comparison (median EPFS, 17 v 9 months; P = .01 by Cox model, adjusting for randomization stratification factors [age, optimal v suboptimal disease, schedule of prior paclitaxel treatment] and baseline CA-125; median PFS, 20 v 12 months; P = .01 by the same COX model), or comparison by baseline CA-125 categories (for baseline CA-125 However, before the EPD criterion is accepted for routine use in clinical trials, its predictive properties must be validated by additional data, preferably prospectively. Ideally, future trials could collect the EPD information as an additional end point. CA-125 data would be collected more regularly. As indicated in Results, a lapse or lack of CA-125 measurements in many patients was the reason for uncertainty regarding the occurrence of either EPD or GCIG CA-125 progression in this analysis. Marker fluctuations related to benign conditions could be ruled out, given that development of a nonspecific inflammatory process and its subsequent treatment (eg, corticosteroids, nonsteroidal anti-inflammatory agents) might influence the serum antigen level. Outlying values such as an isolated nadir could be verified for accuracy. Only then can one avoid the limitations of retrospective data to properly characterize the false-positive rate of the proposed criterion, and refine it as necessary. With an acceptable and validated false-positive rate, the EPD criterion would allow a more accurate evaluation of the biologic (and clinical) effects of a treatment regimen due to its sensitivity, and the objectivity, availability, wide applicability, and cost-effectiveness of CA-125 assays. A second issue relates to the question of the clinical relevance associated with a predictive signal of disease progression earlier than with the current GCIG CA-125 criteria. It is critical to acknowledge the present absence of any prospective randomized phase III trial data demonstrating clinical benefit associated with earlier documentation (by CA-125 or any other criteria) of ovarian cancer disease progression. In fact, based on existing information, one could reasonably argue that detecting disease progression earlier in the absence of cancer-related symptoms will result in increased patient anxiety and the institution of toxic second-line treatment without evidence that such a strategy will lead to an improved outcome, compared with observation until there are more overt signs of the cancer. Two responses can be provided to this concern. First, it is certainly possible that future clinical trials will reveal that earlier initiation of second-line therapy (the specific drugs/strategies to be determined) can influence outcome favorably, and it will be necessary to have a reliable method to document progression from the prior regimen to permit the new program to be initiated. Second, as the detailed description of the follow-up CA-125 levels of several patients (see Results) in this series documents, the natural history of ovarian cancer is a complex process. Although changes in CA-125 levels generally follow a monotone pattern for most patients, it is overly simplistic to assume that all progressing cancers will exhibit the same rate of increase of this serum antigen, or that very slowly progressive increases indicate a false-positive test. Although at present this test will not necessarily inform clinicians if and when treatment should be changed or reinitiated, at least the patient's disease status could be monitored closely when the EPD criterion is met to avoid delay in detecting symptomatic progressive disease.
The author(s) indicated no potential conflicts of interest.
Conception and design: Ping-Yu Liu, David S. Alberts, Maurie Markman Provision of study materials or patients: David S. Alberts, Bradley J. Monk, Maurie Markman Collection and assembly of data: Ping-Yu Liu, David S. Alberts, Bradley J. Monk, Mark Brady, James Moon, Maurie Markman Data analysis and interpretation: Ping-Yu Liu, James Moon, Maurie Markman Manuscript writing: Ping-Yu Liu, Maurie Markman Final approval of manuscript: Ping-Yu Liu, David S. Alberts, Bradley J. Monk, Mark Brady, James Moon, Maurie Markman
We thank the editor and the reviewers for their helpful comments, which greatly clarified the implications of this analysis.
Supported by Public Health Service Cooperative Agreement grants awarded by the National Cancer Institute, Department of Health and Human Services: Grants No. CA38926, CA32102, CA105409, CA58723, and CA13612. Authors' disclosures of potential conflicts of interest and author contributions are found at the end of this article.
1. Vergote I, Rustin GJS, Eisenhauer EA, et al: Re: New guidelines to evaluate the response to treatment in solid tumors [ovarian cancer]. J Natl Cancer Inst 92:1534-1535, 2000 2. Crawford SM, Paul J, Reed NS, et al: The prognostic significance of the CA-125 nadir in patients that achieve a CA-125 response. J Clin Oncol 22:449s, 2004 (suppl; abstr 5001) 3. Markman M, Liu PY, Rothenberg ML, et al: Pretreatment CA-125 and risk of relapse in advanced ovarian cancer. J Clin Oncol 24:1454-1458, 2006 4. Nadal RM, Ojeda BM, Artigas V, et al: Stratification of the normal range of CA-125 after chemotherapy as a predictive factor in carcinoma of the ovary. J Clin Oncol 24:270s, 2006 (suppl; abstr 5059) 5. Tanabe H, Katsumata N, Matsumoto K, et al: CA-125 nadir as a prognostic factor in advanced ovarian carcinoma: A retrospective study of 84 patients achieving clinical CR. J Clin Oncol 24:270s, 2006 (suppl; abstr 5060) 6. Markman M, Liu PY, Wilczynski S, et al: Phase III randomized trial of 12 versus 3 months of maintenance paclitaxel in patients with advanced ovarian cancer after complete response to platinum and paclitaxel-based chemotherapy: A Southwest Oncology Group and Gynecologic Oncology Group trial. J Clin Oncol 21:2460-2465, 2003 7. Santillan A, Garg R, Zahurak ML, et al: Risk of epithelial ovarian cancer recurrence in patients with rising serum CA-125 levels within the normal range. J Clin Oncol 23:9338-9343, 2005 Submitted October 6, 2006; accepted March 23, 2007.
Related Editorial
This article has been cited by other articles:
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|||||||||||
|
|||||||||||
|
|
Copyright © 2007 by the American Society of Clinical Oncology, Online ISSN: 1527-7755. Print ISSN: 0732-183X
|
TOP
ABSTRACT
INTRODUCTION
