Purpose: To describe the data and analyses that led to the U.S. Food and Drug Administration (FDA) approval of ofatumumab (Arzerra, GlaxoSmithKline) for the treatment of patients with chronic lymphocytic leukemia (CLL) refractory to fludarabine and alemtuzumab.
Experimental Design: The FDA reviewed the results of a planned interim analysis of a single-arm trial, enrolling 154 patients with CLL refractory to fludarabine, and a supportive dose-finding, activity-estimating trial in 33 patients with CLL. Patients in the primary efficacy study received ofatumumab weekly for eight doses, then every 4 weeks for an additional four doses; patients in the supportive trial received four weekly doses. In the primary efficacy study, endpoints were objective response rate and response duration.
Results: For regulatory purposes, the primary efficacy population consisted of 59 patients with CLL refractory to fludarabine and alemtuzumab. In this subgroup, the investigator-determined objective response rate was 42% [99% confidence interval (CI), 26–60], with a median duration of response of 6.5 months (95% CI, 5.8–8.3); all were partial responses. The most common adverse reactions in the primary efficacy study were neutropenia, pneumonia, pyrexia, cough, diarrhea, anemia, fatigue, dyspnea, rash, nausea, bronchitis, and upper respiratory tract infections. Infusion reactions occurred in 44% of patients with the first infusion (300 mg) and 29% with the second infusion (2,000 mg). The most common serious adverse reactions were infections, neutropenia, and pyrexia.
Conclusions: On October 26, 2009, the FDA granted accelerated approval to ofatumumab for the treatment of patients with CLL refractory to fludarabine and alemtuzumab, on the basis of demonstration of durable tumor shrinkage. Clin Cancer Res; 16(17); 4331–8. ©2010 AACR.
Chronic lymphocytic leukemia (CLL) is a clonal disorder of B cells with a variable clinical course. Median survival is longer than 10 years4; however, reported median survival is only 2 to 3 years for patients with high-risk disease (Rai category III or IV or Binet stage 3; ref. 1). Reported survival in CLL patients whose disease is refractory to multiple drugs is shorter; in one report, median survival was 8 months in patients with CLL refractory to fludarabine and alemtuzumab (2).
In the past decade, regular U.S. Food and Drug Administration (FDA) approval for the treatment of patients with CLL has been based on a clinically meaningful prolongation of progression-free survival (PFS) accompanied by an acceptable safety profile; whereas accelerated approval has been granted on the basis of durable overall response rates (ORR) in patients with CLL, which has progressed following available therapy. In order to be considered for accelerated approval, a drug that shows an effect on a surrogate endpoint that is reasonably likely to predict clinical benefit must treat a serious or life-threatening disease or condition and show an advantage over available therapy (3). Additionally, the commercial manufacturer must commit to study the drug further to verify and characterize the predicted clinical benefit (3).
Alemtuzumab received accelerated approval in 2001 and regular approval in 2007. Accelerated approval was based on the results of three single-arm studies enrolling 149 patients with CLL and progressive disease following alkylating agents and fludarabine. The rates of durable objective responses (ORR) in the three studies ranged from 21 to 33% (4). Regular approval was based on superior PFS [HR 0.58 (95% confidence interval [CI], 0.43–0.77), P < 0.0001 stratified log-rank test] in a randomized active-controlled study comparing alemtuzumab to chlorambucil in previously untreated patients with CLL (4). Alemtuzumab also showed an improvement in ORR (83% versus 55%) and complete response rates (24% versus 2%) compared with chlorambucil (4).
The FDA approved bendamustine in 2008 on the basis of superior PFS [HR 0.27 (95% CI, 0.17-0.43), P < 0.0001] in a randomized active-controlled study comparing bendamustine to chlorambucil in previously untreated patients with CLL. In the study, bendamustine also showed an improvement in ORR (59% versus 26%) and complete response rates (8% versus <1%) compared with chlorambucil (5).
In addition to bendamustine and alemtuzumab, the following drugs are approved for the treatment of patients with CLL: fludarabine, cyclophosphamide, and chlorambucil. These drugs were approved on the basis of durable objective tumor responses prior to the codification of the accelerated approval regulations.
The Biologics License Application (BLA) for ofatumumab was submitted on January 30, 2009. The application contained the interim results of a single-arm, fixed-dose, multicenter trial (Hx-CD20-406), and the final results of a multicenter, sequential dose-escalating cohort and activity-estimating trial (Hx-CD20-402), which used a shorter treatment duration (4 doses versus 12 doses) than the primary efficacy study. Both trials enrolled adults with CLL who had relapsed following, or were refractory to, one or more standard treatments. The BLA was supported by safety information from several trials enrolling 151 patients with follicular lymphoma or diffuse large B-cell lymphoma, 282 patients with rheumatoid arthritis or chronic obstructive pulmonary disease, and 28 patients with CLL who received ofatumumab in combination with chemotherapy. The primary efficacy trial used the highest dose and longest duration of ofatumumab in the clinical safety database.
Chemistry, Manufacturing, and Controls
Ofatumumab is an immunoglobulin G1 (IgG1) kappa human monoclonal antibody with a molecular weight of approximately 149 kDa (6). The antibody was generated via transgenic mouse and hybridoma technology and is produced in a recombinant murine cell line (NS0) using standard mammalian cell cultivation and purification technologies (6). It is supplied as a sterile, colorless, preservative-free liquid concentrate at a concentration of 20 mg/mL in a 5 mL volume of solution for intravenous infusion in a 10 mL single use vial (100 mg/vial; ref. 6).
Nonclinical Pharmacology and Toxicology
Ofatumumab binds specifically to the CD20 molecule, which is expressed on normal B lymphocytes (pre-B– to mature B–lymphocyte) and on B-cell CLL tumor cells. The CD20 molecule is not shed from the cell surface and is not internalized following antibody binding (6). The Fab domain of ofatumumab binds to the CD20 molecule, and the Fc domain mediates immune effector functions to result in B-cell lysis in vitro (6). Possible mechanisms of action for ofatumumab include complement-dependent cytotoxicity and antibody-dependent cell-mediated cytotoxicity.
Nonclinical data submitted in support of the application included tissue-binding studies in human and cynomolgus monkey tissue panels, evaluation of in vitro pharmacologic activity (receptor binding, lysis of CD20-expressing cells), in vivo pharmacologic activity (antitumor activity in severe combined immunodeficient mice bearing human tumor xenografts, characterization of B-cell depletion in treated monkeys), pharmacokinetic studies in cynomolgus monkeys, and subacute (4-week) and chronic (7-month) toxicology studies in cynomolgus monkeys. The major ofatumumab toxicities observed in nonclinical studies were extensions of its expected pharmacologic activity, that is, the severe and prolonged depletion of B lymphocytes both in the circulation and in the major lymphoid organs (follicular germ centers of the spleen, cortical lymph nodes, and Peyer's patches in the gastrointestinal tract). Other toxicities observed in the 7-month repeat-dose cynomolgus monkey toxicity study included sporadic, nonsevere infusion reactions, hematologic changes including B-cell depletion, mild hemolytic anemia, and changes in neo- and recall-antigen responses.
Pharmacokinetic data were obtained from 146 patients with refractory CLL who received a 300-mg initial dose followed by seven weekly and four monthly infusions of 2,000 mg (6). The Cmax and AUC(0−∞) after the eighth infusion in study Hx-CD20-406 were approximately 40% and 60% higher than after the fourth infusion in study Hx-CD20-402 (6). The mean volume of distribution at steady-state (Vss) values ranged from 1.7 to 5.1 L (6). Ofatumumab is eliminated through both a target-independent route and a B-cell–mediated route (6). Ofatumumab exhibited dose-dependent clearance in the dose range of 100 to 2,000 mg (6). Because of the depletion of B cells, the clearance of ofatumumab decreased substantially after subsequent infusions compared with the first infusion (6). The mean clearance between the 4th and 12th infusions was approximately 0.01 L/h and exhibited large intersubject variability with CV% greater than 50% (6). The mean t1/2 between the 4th and 12th infusions was approximately 14 days (range: 2.3 to 61.5 days; ref. 6).
No dosage adjustment was determined to be necessary on the basis of body weight, age, gender, or renal impairment. Patient age (ranging from 21 to 86 years) and creatinine clearance at baseline (ranging from 33 to 287 mL/min) did not have a clinically important effect on ofatumumab pharmacokinetics (6). Gender did not influence ofatumumab pharmacokinetics after correcting for body weight differences. Although volume of distribution and clearance increased with body weight (6), it did not explain intersubject variation for the clinical endpoint.
In patients with CLL refractory to fludarabine and alemtuzumab, the median decrease in circulating CD19-positive B cells was 91% (n = 50) with the 8th infusion and 85% (n = 32) with the 12th infusion (6). The time to recovery of lymphocytes, including CD19-positive B cells, to normal levels could not be determined using the data submitted to the BLA.
Materials and methods
The major efficacy (Hx-CD20-406) and dose-finding (Hx-CD20-402) trials were conducted by Genmab A/S. Study Hx-CD20-402 was a multicenter, sequential dose-escalating and activity-estimating trial that enrolled a total of 33 patients with previously treated CLL. Three dose cohorts were studied with an expansion of the highest dose cohort to further evaluate antitumor activity. The treatment regimen was limited to four weekly doses as follows: level 1 received 100 mg week 1 and 500 mg weeks 2 to 4 (n = 3); level 2 received 300 mg week 1 and 1,000 mg on weeks 2 to 4 (n = 3); and level 3 received 500 mg on week 1 and 2,000 mg weeks 2 to 4 (n = 27).
Hx-CD20-406 was an open-label, multicenter, international, single-arm, fixed-dose trial. Eligibility was limited to patients 18 years of age or older, Eastern Cooperative Oncology Group (ECOG) performance status (PS) 0 to 2, with CLL requiring treatment as defined in the NCI Working Group (1996 NCIWG) guidelines (7). All patients were required to be refractory to fludarabine (minimum of two cycles), and to be refractory to alemtuzumab [double refractory (DR) subgroup], or to be refractory to fludarabine and have at least one lymph node >5 cm [bulky fludarabine refractory (BFR) subgroup]. A total of 16 patients met entry criteria for earlier versions of the study but could not be considered DR or BFR. Drug refractoriness was defined as the failure to achieve at least a partial response to, or disease progression within 6 months of, the last dose of fludarabine or alemtuzumab. The primary efficacy population was a prespecified subgroup of those enrolled in Hx-CD20-406, the fludarabine- and alemtuzumab-refractory subgroup, based on presubmission discussions with the FDA.
Ofatumumab was administered as an intravenous infusion in 1,000 mL of 0.9% sodium chloride injection, USP, weekly for the first eight doses (doses 1 to 8), then every 4 weeks for the remaining four doses (doses 9 to 12). The first dose of ofatumumab was reduced (300 mg) to reduce the risk of serious infusion reactions; all subsequent doses were 2,000 mg. Premedication with an oral or intravenous antihistamine, oral acetaminophen, and an intravenous corticosteroid (100 mg prednisolone or equivalent) was administered prior to each dose; the dose of corticosteroid administered prior to doses 3 to 8 and 10 to 12 could be reduced on the basis of the absence of infusional toxicity on preceding doses.
Ofatumumab dosing was terminated early for a critical adverse event, pregnancy, administration of alternative anticancer therapy, withdrawal of consent, or investigator discretion. Critical adverse events included the following: occurrence of a treatment-related grade 4 or greater adverse event on any noninfusion day (excluding infections); a grade 3 or greater treatment-related adverse event that prevented resumption of an ofatumumab infusion; second occurrence of grade 3 or greater bronchospasm during an infusion; or the third occurrence of the same adverse event of grade 3 or greater in severity during an infusion. Toxicity was managed by reducing the rate of infusion, suspension of an infusion for management of infusional toxicity, or by withholding the ofatumumab dose for other toxicities.
Protocol-specified monitoring included assessment of CLL status at baseline, every 4 weeks until week 28, then every 3 months until month 24. CLL disease-status assessments consisted of measurements of lymph nodes, liver, and spleen by physical examination; complete blood counts including hemoglobin, platelet, neutrophil, and lymphocyte counts; and queries for constitutional symptoms. Bone marrow aspiration and biopsy and computed tomography (CT) scan of the neck, chest, abdomen, and pelvis were obtained at baseline and for confirmation of a complete response (CR). All adverse events were collected up to 24 months following study entry; serious adverse events, resulting in hospitalization or death, and peripheral B-lymphocyte counts were collected for 48 months or until initiation of alternative anti-CLL therapy, whichever occurred earlier.
The primary endpoint was objective response rate (ORR), based on the 1996 NCIWG criteria for response, occurring at any time between the first dose of ofatumumab through week 24, as determined by an Independent Endpoints Review Committee (IRC). The IRC was composed of five independent board-certified oncologists or hematologists experienced with CLL. The committee was to be provided with clinical data and determine the level of response and when progression occurred according to NCIWG guidelines. CT scans were not included in the response determinations. An independent radiology review was conducted by an independent imaging core laboratory (ICON medical imaging). A single board-certified radiologist reviewed all imaging data to: (1) confirm complete remissions and (2) verify the presence of bulky lymphadenopathy at study entry.
The protocol-specified criteria for determination of ORR (complete plus partial responses [PR]) were based on the 1996 NCIWG guidelines (7). The criteria for a PR required the following: ≥50% decrease in peripheral blood lymphocyte counts; ≥50% reduction in lymphadenopathy [sum of the perpendicular products of all palpable lymph nodes (the largest lymph node in each area was to be measured)]; ≥50% reduction in hepatomegaly or splenomegaly from baseline measurements if there was hepatomegaly or splenomegaly at baseline; and at least one of the following: polymorphonuclear leukocyte count ≥1,500/mcL or 50% improvement over baseline; platelet count >100,000/mcL or 50% improvement over baseline; or hemoglobin >11 gm/dL or 50% improvement over baseline without transfusions with a response duration of at least 56 days. The criteria for a CR required the absence of lymphadenopathy, hepatomegaly, and splenomegaly by physical examination; the absence of constitutional symptoms; normal hematological counts with lymphocytes less than 4,000/mcL for at least 56 days; and a bone marrow aspirate (2 months after CR criteria were met) showing the absence of lymphoid nodules, <30% lymphocytes, and normocellularity for age. The protocol also required that a CT scan be done 8 weeks after a patient fulfilled the requirements of a CR (to confirm the CR).
Additional endpoints were duration of response (DOR), PFS, time-to-next CLL therapy, overall survival, and reduction in tumor size (the percentage change in the sum of the products of the diameters). The DOR was defined as the time from the initial response to disease progression, or death. For the analysis of DOR, the following scenarios were censored: no progression at the end of the trial; treatment discontinued for undocumented progression, toxicity, or other reasons; new anticancer therapy started; and death or progression after two or more consecutive missed visits.
The planned study sample size of 100 patients per treatment subgroup (DR and BFR) was based on the assumption of 30% ORR in the DR subgroup and the probability that the exact 2-sided 99% CI would exclude a response rate of less than 15% with 63% power based on 66 patients in the interim analysis and 92% power based on 100 patients in the final analysis.
The initial activity-estimating trial, Hx-CD20-402, evaluated 27 patients at the highest dose cohort. There was a 42% PR rate, with a median duration of response of 16 weeks across the entire study (n = 33). One response was reported in the 500 mg (n = 3) cohort, none reported in the 1,000 mg (n = 3) cohort, and 13 responses reported among the 27 patients enrolled at the 2,000-mg dose cohort. Because of the relatively modest duration of responses observed with four weekly doses, the duration of treatment was increased in the Hx-CD20-406 trial.
The Hx-CD20-406 trial enrolled 154 patients at the time of the planned interim analysis. The IRC conducted an audit of patient records and case report forms to confirm investigator-assigned patient subgroup classification; the IRC review identified 59 patients as DR, 79 patients as BFR, and 16 patients as “other.” This review focuses on the efficacy results from the IRC-confirmed DR group, the patient population that satisfied the regulatory definition of unmet medical need.
The overall population in study Hx-CD20-406 was heavily pretreated. In the DR group (n = 59), the median number of prior therapies was 5 and the range was 1 to 14. Two patients in the DR group received one prior therapy; both of these patients received fludarabine combined with alemtuzumab. Baseline characteristics and prior treatment history are summarized for the DR subgroup and for the overall study population in Tables 1 and 2, respectively. Nearly all (97%) patients enrolled in study Hx-CD20-406 were white, and most were men (72%). Median time in years from original diagnosis to enrollment into study Hx-CD20-406 was 6.3 years. ZAP70 was not assessed at baseline. Approximately one quarter (26%) of the study population was accrued from U.S. sites.
Table 2 shows that nearly all patients in the DR group, as well as the overall study population, received an alkylating agent–containing regimen. More than 50% of the patients in study Hx-CD20-406 received prior rituximab; however, data were not collected to determine whether these patients were rituximab refractory.
The magnitude of the treatment effect on ORR, as determined by the individual IRC members, IRC consensus determination, investigators, and FDA, differed. The ORR and DOR determined by the investigator, the IRC consensus finding, and FDA are summarized in Table 3. The estimated DOR of response in the DR population was 6.5 months.
Unexpected findings in this trial were the high rate of discordance in tumor-response assessment between the individual IRC members and the higher ORR as determined by the IRC as compared with that determined by the investigators. Investigator assessments from study Hx-CD20-406 were derived from assessments of response at individual visits and not from an overall response assessment provided by the investigators. Fifty-eight percent of 59 patients in the DR and of the 79 patients in the BFR subgroups underwent adjudication by a third IRC member. Adjudication was required when there was disagreement between the initial IRC reviewers about response assessment, date of onset of the response, or date of progression. The high rate of disagreement was unexpected, given that all IRC members used the same information for lymphocyte measurements, hematology laboratory values, and investigator-recorded liver, spleen, and lymph node measurements. There was insufficient information captured during the IRC review to allow the FDA to determine the source of the disagreements between IRC members.
These differences seem to have been related to individual interpretation of the 1996 NCIWG criteria as applied to the same set of data. The 1996 NCIWG response criteria were written such that a strict application of the criteria would result in few responders, even in a drug with notable activity. For example, the criteria for disease progression could be met with new lymphadenopathy of relatively small size, even if transiently enlarged in a patient with an intercurrent infection.
The FDA clinical reviewer's determination of ORR, based on data captured in case report forms and study data sets, yielded an ORR in the DR subgroup that was similar to that of the investigators and lower than that of the IRC. The FDA clinical reviewer did not use the strictest interpretation of the 1996 NCIWG criteria under which the detection of any new node (for example, a 1 × 1–cm lymph node that regressed at the next visit) would designate a progression event.
Study Hx-CD20-406 was the only study submitted to the BLA that evaluated ofatumumab at the recommended doses and schedule. Because the data primarily were derived from an uncontrolled study, only a descriptive analysis of safety could be done. In study Hx-CD20-406, 90% of patients received at least 8 infusions of ofatumumab, and 55% received 12 infusions. The most common reasons for stopping treatment was disease progression and adverse reactions, mostly due to infections.
The most common adverse reactions (≥10%) in study Hx-CD20-406 were neutropenia, pneumonia, pyrexia, cough, diarrhea, anemia, fatigue, dyspnea, rash, nausea, bronchitis, and upper respiratory tract infections. Table 4 shows adverse reactions that occurred in ≥5% of the total study population. The incidence of neutropenia as an adverse reaction is not described in the table because the laboratory data gave a more accurate description of neutropenia occurring after ofatumumab treatment.
Infusion reactions, including sequelae of such reactions, occurred frequently during the ofatumumab development program (all oncology and nononcology studies). Infusion reactions were less common after the first two infusions. Serious infusion-related reactions or sequelae included bronchospasm, dyspnea, laryngeal edema, pulmonary edema, flushing, hypertension, hypotension, syncope, cardiac ischemia and/or infarction, back pain, abdominal pain, pyrexia, rash, urticaria, and angioedema. In a separate study of patients with moderate-to-severe chronic obstructive pulmonary disease, two of five patients developed grade 3 bronchospasm during an infusion. In study HxCD20-406, infusion reactions occurred in 44% of patients on the day of the first infusion (300 mg), and 29% on the day of the second infusion (2,000 mg).
Infections frequently occurred after treatment with ofatumumab. In study Hx-CD20-406, 70% of patients experienced bacterial, viral, or fungal infections. A total of 29% of the study population experienced ≥grade 3 infections, of which 12% were fatal. The proportion of fatal infections in the DR group was 17%. One patient with CLL died of progressive multifocal leukoencephalopathy after receiving ofatumumab treatment in study Hx-CD20-406. One patient with rheumatoid arthritis died of fulminant hepatitis B infection after receiving ofatumumab in a nononcology study.
Neutropenia frequently occurred or was exacerbated following ofatumumab treatment. Of 108 patients with normal baseline neutrophil counts at baseline, 42% developed grade 3 or greater neutropenia. Nineteen (18%) developed grade 4 neutropenia. Some patients experienced new onset grade 4 neutropenia for more than 2 weeks in duration.
The FDA granted subpart E (accelerated) approval under 21 CFR 601.41 for ofatumumab on October 26, 2009. Accelerated approval may be granted only for a drug that treats a serious or life-threatening illness and provides meaningful therapeutic benefit to patients over existing treatments (i.e., fulfills an unmet medical need), on the basis of the demonstration of an effect on a surrogate endpoint that is reasonably likely to predict clinical benefit. The FDA concluded that the subpopulation of patients studied in Hx-CD20-406 who were refractory to fludarabine and alemtuzumab, more than 90% of whom also received prior therapy that included an alkylating agent and who received a median of five prior treatments prior to enrollment in the Hx-CD20-406 study, was a patient population with an unmet medical need. The ORR effect size in the DR population of 42%, with an estimated duration of response of 6.5 months supported by antitumor activity in other patients with CLL, was considered reasonably likely to predict clinical benefit both by the FDA and by the majority of the May 2009 Oncology Drugs Advisory Committee (10-to-3 vote).
For the product label, the FDA limited the description of efficacy to the subgroup that was refractory to fludarabine and alemtuzumab, as this was the population that met the regulatory definition of unmet medical as required for accelerated approval. For accelerated approval, drugs must provide meaningful therapeutic benefit to patients over existing treatments. In settings where existing treatments are available, randomized studies are necessary to grant accelerated approval on the basis of an effect on a surrogate endpoint. The randomized studies should show that the new drug provides meaningful therapeutic benefit over an existing treatment. This comparative effectiveness standard exists for accelerated approval but does not exist for regular FDA approval of a drug or biological drug.
The applicant did not satisfactorily establish that patients with BFR disease met the regulatory definition of unmet medical need. The Hx-CD20-406 protocol required that patients with bulky disease be refractory to fludarabine and thus may have received only one previous treatment regimen. However, responses to alemtuzumab were reported in 76% of 33 previously untreated patients with lymph nodes ≥5 cm (CAM307 study; ref. 8). As stated by the FDA in the May 29, 2009 advisory committee meeting, the FDA could not determine whether ofatumumab is better than alemtuzumab in the absence of a randomized comparative trial.
The FDA used the investigator-determined ORR and DOR for regulatory decision making. The FDA elected not to use the IRC determination for several reasons, including the high rate of disagreement across reviewers, the lack of direct evaluation of tumor measurements based on lack of serial radiographs for independent analysis, the relatively minor contribution of hematologic laboratory data to the determination of ORR, and the FDA's independent review of the case reports forms and supporting laboratory data.
The magnitude of the ORR in study Hx-CD20-406 could not be reproduced by multiple groups (investigators, IRC members, and FDA) given the same set of information. This lack of consistency seemed to be due to variability in interpretation and application of the 1996 NCIWG criteria for response assessments. In order to have consistent application, tumor-response criteria should ideally be unambiguous such that given the same data, similar conclusions will be drawn by all response assessors. Such consistent application of the response criteria was not observed during the review of study Hx-CD20-406. Criteria that allow for such variability in interpretation are generally appropriate for the overall practice of medicine; however, the evidentiary standard required for regulatory decision making is generally higher than those employed for clinical decision making, or for informing the overall practice of medicine.
Since the Hx-CD20-406 study was started, the NCIWG published new (2008) criteria that now recommend the use of CT scans in the response assessments of patients with CLL for clinical trials (9). On the basis of these revised criteria and advice from the May 2009 Oncology Drugs Advisory Committee (ODAC), the FDA will expect that serial CT scans be obtained in new studies in order to make labeling claims for ORR in patients with advanced CLL. The FDA notes, however, that certain aspects of the revised 2008 criteria may still be problematic. For example, if one lymph node increases in size from 1 cm to 1.5 cm during the first 2 months of a response, a patient may not be classified as a responder, even if the node resolves in size at the next visit and all other nodes have decreased in size. Additionally, a single new node of 1 cm in diameter during the first 2 months of a response can designate the patient as a nonresponder.
Occurrences such as these were not considered to nullify a response in study Hx-CD20-406, as clinical judgment was allowed to override the explicit rules of the NCIWG. Optimally, the NCIWG criteria should be modified for use in clinical trials so that the rules explicitly state what deviations from the criteria are acceptable for the determination of response or progression. Such codification of the rules will permit a more accurate description of the treatment effect of a drug. The FDA encourages sponsors to discuss such deviations from the response criteria prior to initiating a clinical study intended to support labeling claims based on PFS or ORR in patients with CLL.
The FDA acknowledges that uncertainty exists in the ability of a surrogate endpoint to predict direct clinical benefit, and therefore, under 21 CFR 601.41, additional clinical studies are required to verify and describe the clinical benefit of the drug. GlaxoSmithKline has committed to conduct a confirmatory randomized trial of ofatumumab plus chlorambucil versus chlorambucil for the first-line treatment of elderly or fragile patients with CLL who are unable to tolerate fludarabine-based chemotherapy regimens. The study is designed to show a clinically meaningful improvement in PFS. Regular FDA approval for the treatment of patients with CLL has been based on a clinically meaningful prolongation of PFS accompanied by an acceptable safety profile. Superior PFS may represent a direct clinical benefit if there is a large treatment effect and the risks are acceptable.5 In hematologic malignancies, the clinical significance of a delay in disease progression is clearer because progression is generally associated with increased transfusion requirements and increased risk of infections. In patients with indolent hematologic malignancies, in which median survival is measured in years, demonstration of effects on overall survival may not be as practical as an approval endpoint; however, evidence of no apparent detrimental effects on survival is generally required as an assessment of product safety.
Enrollment of the confirmatory GlaxoSmithKline trial will be limited to patients deemed inappropriate for fludarabine-based treatment. In addition to confirming the clinical benefit of ofatumumab, a randomized confirmatory trial will allow for a more accurate characterization of adverse reactions caused by ofatumumab. Characterization of infections was especially problematic because the background rate of severe and fatal infections in heavily treated CLL patients is high (10). Because of the high background rate of infections in this patient population and the absence of an internal control, it was not possible to determine the additional risk of infection posed by the administration of ofatumumab. However, ofatumumab-induced neutropenia may increase the risk of life-threatening infections in this patient population.
Disclosure of Potential Conflicts of Interest
No potential conflicts of interest were disclosed.
↵5Guidance for Industry, Clinical Trail Endpoints for the Approval of Cancer Drugs and Biologics. Rockville (MD): U.S. Department of Health and Human Services, Food and Drug Administration; May 2007 [cited 2010 May 24]. Available from: http://www.fda.gov/downloads/Drugs/GuidanceComplianceRegulatoryInformation/Guidances/ucm071590.pdf.
- Received March 29, 2010.
- Revision received June 11, 2010.
- Accepted June 14, 2010.