Issue Analysis Summary on Maintenance Therapy in Advanced Non-Small Cell Lung Cancer

Help on accessing alternative formats, such as Portable Document Format (PDF), Microsoft Word and PowerPoint (PPT) files, can be obtained in the alternate format help section.

Contact: Bureau of Policy, Science and International Programs Enquiries

Martha Harczy MD Msc FRCPC
Therapeutic Products Directorate
Bureau of Metabolism, Oncology and Reproductive Sciences (BMORS)
Division of Oncology 1
April 15, 2010

Contents

• 1. Issue
• 2. Purpose/Objective
• 3. Background
• 4. Issue Analysis
• 5. Consultations
• 6. Recommendation
• 7. References

1. Issue

Health Canada is seeking advice from its Scientific Advisory Committee on Oncology Therapies on clinical merits of "maintenance therapy" in advanced Non-Small-Cell Lung Cancer (NSCLC) in patients who did not progress during first line combination chemotherapy. Maintenance therapy in this context is defined as the addition of further systemic therapy in patients who have responded or achieved at least stable disease following first line platinum based chemotherapy¹⁸.

2. Purpose/Objective

To assist Health Canada in regulatory decision making for products seeking regulatory approval for "maintenance therapy" of advanced (stage III/IV) NSCLC. Maintenance therapy is yet to be adopted as standard modality in this disease.

3. Background

Advanced NSCLC is an incurable disease. The term "advanced" includes stage III locally advanced disease which is not amenable to potentially curative therapies (stage IIIb non-resectable disease) and stage IV (metastatic) or recurrent disease. The median overall survival (OS) is 8-11 months¹ despite adequate effective combination chemotherapy. A landmark meta-analysis comparing supportive care with supportive care plus chemotherapy resulted in 10% improvement in survival at one year and 2 months in median survival² .

At diagnosis of advanced NSCLC first line platinum based combination chemotherapy is the standard of care in Canada for patients who are suitable to such therapy. The most common combinations used in Canadian cancer clinics are: cisplatin+gemcitabine; cisplatin+vinorelbine; carboplatin+taxol. These combinations are given in 3-4 weeks cycles. The intent of therapy is palliation (that is [i.e.] symptom control, maintenance of quality of life, prolong survival). The optimal duration of this first line chemotherapy has been addressed in several phase III studies³, ¹⁰, ¹¹. In these studies there were no additional benefit of first line therapy longer than 4-6 cycles on quality of life or survival (OS). Based on those studies and the lack of evidence that "more is better" the guidelines of American Society of Clinical Oncology adopted since 2003 that no more than 4 cycles for patients who are not responding and no more than 6 cycles for responding patients should be given. Similar recommendations were made in the NCCN guidelines with regard to the number of cycles of first line therapy. Thus in Canada and in North America following the diagnosis of advanced NSCLC 4-6 cycles of platinum based combination chemotherapy is the standard first line therapy. First line combination chemotherapy and induction therapy are synonymous terms. This is followed by close observation for recurrence or progression. Should progression occur, approved second line therapies (docetaxel, pemetrexed, erlotinib) are available in Canada and generally are applied as single agents.

Approval of maintenance therapy would be based on a demonstration of clinical benefit of continuation of a "maintenance" (chemotherapy or targeted agent) therapy which was not used in the first line setting.

Several phase III studies addressed the issue of maintaining active therapy after completion of 4-6 cycles of first line therapy. These can be grouped in 3 major different design categories.

1. First line combination chemotherapy for 4-6 cycles followed by best supportive care (BSC) until disease progression versus the same first line combination extended until disease progression;
2. Defined cycles of first line chemotherapy versus a longer defined cycles of the same chemotherapy;
3. Defined cycles of first line chemotherapy followed by BSC versus the same first line chemotherapy followed by additional cycles of alternative chemo-, or targeted therapy and BSC until progression.

A recent meta-analysis by Soon et al.⁵ addressed the benefit of maintenance chemotherapy on OS as primary outcome and on the following secondary objectives: progression free survival (PFS), adverse events (AE) and health related quality of life (HRQL).

In this meta-analysis 13 randomized clinical trials (RCT) were included, representing 3027 patients. A substantial improvement in PFS was found (HR:0.75;95% CI,0.69 to 0.81;p<.00001) along with a modest improvement in OS (HR:0.92;95% CI 0.86 to 0.99;p=.03). The effects of extended chemotherapy on OS became statistically significant with inclusion of one large trial⁶.

Subgroup analysis of studies with the above outlined 3 different designs revealed that the effect on PFS was greater in those RCT-s which extended first line chemotherapy with "third generation" regimens (gemcitabine, docetaxel, pemetrexed, erlotinib, gefitinib).

Refer to: OS analysis of extended (maintenance) chemotherapy by meta-analysis of Soon et al.⁵ J. Clin. Oncol 2009;27(20:3277-83).

Comments pertaining to this meta-analysis :

1. Randomization heterogeneity was observed in these trials of Soon meta-analysis
   1. Some studies randomized patients up-front of first line therapy and carried through all patients including progressive diseases³, ¹¹, ¹²;
   2. Other studies randomized a pre-selected population only those who were responding to first line therapy, fit (ECOG:0-1) and willing ⁶, ⁴,¹³, ¹⁴,¹⁵. This population may have a better survival than the non selected patients with advanced NSCLC.
   3. Some used minimization techniques⁶ to minimize prognostic imbalances (Pocock and Simon method).
2. Design heterogeneity: The three major designs as it was outlined above.
3. Overall survival reporting heterogeneity:
   1. Some studies reporting OS from the beginning of first line therapy
   2. Other studies reporting OS following randomization (after first line) for maintenance therapy.
4. Patient enrollment heterogeneity:
   1. Some studies enrolled all patients with at least stable disease (complete remission [CR],partial remission [PR],stable disease [SD]);
   2. Some others excluded SD¹⁴
5. Statistical heterogeneity was taken into account and some correctional measures were applied.

4. Issue Analysis

1. When PFS is used as a primary endpoint in pivotal studies and OS as a secondary endpoint, the issues are as follows:
   1. PFS has not been validated either in first line treatment or as maintenance therapy of advanced NSCLC. No detailed meta-analysis is available on surrogacy of PFS for OS with first line platinum based combination in the literature ⁷.
   2. PFS data are follow up schedule dependent²⁰. Most of the studies in the Soon meta-analysis employed 6-8 weekly evaluation of disease by investigator and imaging. Any deviation from this kind of maintenance follow up schedule would influence PFS data and would preclude cross trial comparability of PFS.
   3. Most studies of the Soon⁵ meta-analysis for maintenance therapy for advanced NSCLC used OS as primary endpoint. Not all the trials in the meta-analysis reported PFS.
   4. OS is a more appropriate primary endpoint ⁸, ⁹, ¹⁷, ¹⁹, in advanced NSCLC. This is so because median survival of this disease is relatively short.
   5. When PFS is used as an endpoint it neglects the impact of therapy on subsequent effective therapies after progression. Prolonged side effect of maintenance therapy may prohibit the use of subsequent standard second line therapy. It is interesting to see a)what percentage of patients were able to receive subsequent therapy and b)whether they received the standard second line therapies or not.
2. There are common methodological issues with PFS:
   1. A recent publication on PFS by Dancey et al¹⁵states"..baseline anatomic scanning to detect the presence of disease should include all areas likely to be the sites of metastases." Not all of the metastatic predilection sites of NSCLC were prospectively checked in many trials of the Soon meta-analysis. It is common that only "target sites" are being followed with imaging, thus new diseases appearing on non-target sites are ignored until they become symptomatic. Although this is a cost effective clinical practice a clinical trial which uses an event based primary endpoint (PFS) should have appropriate staging investigations and prospective follow up. For example events of bone metastases were not documented at baseline for every patients and not followed prospectively. The ASCO 2003 guidelines recommend FDG-PET scan or bone scan to rule in or out distant metastasis. Bone scan was only done on those patients with known bone metastasis. Thus a significant lead time bias was introduced in detection of new asymptomatic bone metastases until they become symptomatic or cause fractures. Thus these PFS results can not be compared across historical studies from the meta-analysis which used prospective bone scan evaluations for new metastases as events. One may say: the study was randomised, that this bias is present on both arms, however not all patients develop bone metastases and those factors are unknown which render one patient to develop bone metastasis and other patients exempt from it.
   2. In one of the studies physical examination by investigator was done at baseline, 12 weeks and at progression. This prolonged follow up time may artificially have extended PFS by picking up metastases late which are detectable on physical examination (supra-clavicular, sub-dermal metastases etc.). Such prolonged follow up of actively treated patients is not the clinical practice in Canada. Patients, who are being treated actively are regularly followed up 4-6 weekly. Thus leaving the study results not applicable to Canadian practice.

5. Consultations

Pertaining to maintenance therapy for advanced NSCLC the following general questions are posed to SAC-OT.

1. The cited literature ⁸, ⁹, ¹⁷ recommends overall survival as primary endpoint for randomized clinical trials aiming at advanced NSCLC. This should apply to maintenance therapy¹⁹. Would the SAC-OT concur with that opinion?
2. If so, what is a clinically meaningful OS advantage that the SAC-OT would consider for maintenance therapy?
3. What patient-follow up schedule would the SAC-OT recommend in future clinical trials aiming at maintenance therapy for NSCLC which would be relevant to Canadian practice?
4. Should all the common metastatic predilection sites (lung, bone, liver, adrenal glands, brain) of NSCLC be scanned at baseline and prospectively followed up for all the patients taking part in a clinical trial?
5. For an incurable disease such as advanced NSCLC where the intent of therapy is palliation, is it acceptable that HRQL is either impaired on therapy or not improved?

6. Recommendation

The answers of those questions will serve as the SAC-OT recommendations.

7. References