CENTER FOR DRUG EVALUATION AND
`
`RESEARCH
`
`APPLICA TION NUMBER:
`
`NDA 22-249
`
`CROSS DISCIPLINE TEAM LEADER REVIEW .
`
`

`

`Cross-Disci n line Team Leader Review
`
`Sub'ect
`NDA# Sun '
`Prorieta
`/ Established
`
`Dosage forms lstrength
`
`
` Recommended:
`
`Proposed Indication(s)
`
`SAN names
`
`March 20' , 2008
`Amna Ibrahim MD
`Cross-Disci line Team Leader Review
`22249 S-000
`Bendamustine Treanda
`
`IOO-mg Vials of bendamustine HCL as white to off-
`white 1 o hilized . owder
`
`For the treatment of patients with chronic lymphocytic
`leukemia (CLL). Efficacy relative to first line therapies
`other than Chlorambucil has not been established.
`A uroval
`
`
`
`
`
`This an amended CDTL review. The additions are provided in italics in this review and are based on
`the new information.
`
`1 Introduction
`
`A single, open-label, multicenter, randomized trial has been submitted as the major trial to support the
`approval of bendamustine (Treanda®) for the treatment of patients with CLL. The clinical team
`recommends approval of this NDA on the basis of an improvement in Overall Response Rates (OR)
`and Progression-free Survival (PFS).
`
`In this study, Bendamustine was compared to Chlorambucil as a comparator in a treatment-naive
`population. The choice of comparator was influenced by the drugs approved in Europe for this
`indication, where the trial was conducted. Fludarabine, one of the most active drugs for this disease
`was approved only for second-line use. FDA does not require the use of a standard of care in a
`randomized study.
`
`Bendamustine is a bifunctional nitrogen mustard derivative. Nitrogen mustard and its derivatives are
`alkylating agents which dissociate into electrophilic alkyl groups. These groups form covalent bonds
`with electron-rich nucleophilic moieties. The bifunctional covalent linkage can lead to cell death Via
`several pathways. The exact mechanism of action of bendamustine remains unknown.
`
`CLL is a disease that mainly affects the older population, the median age being 72 at diagnosis. Over
`the past few decades, there has been little progress in prolonging survival of patients with CLL, and it
`remains an incurable disorder. Because the patients generally have a good long term prognosis and
`treatment does not change the outcome of disease, a “watch and wait” approach is often used before
`initiation of treatment. Factors which generally prompt the initiation of therapy include the presence of
`disease-related symptoms, massive and/or progressive lymphadenopathy or hepatosplenomegaly, bone
`marrow failure, or recurrent infections. The lymphocyte doubling time should be considered in the total
`clinical picture but not used as the primary criterion. The routine availability of peripheral blood
`lymphocyte immunophenotyping has facilitated the diagnosis of CLL in patients with a monoclonal
`lymphocytosis. Three main phenotypic features define B—CLL: the predominant population shares B—
`cell markers (CD19, CD20, and CD23) with the CD5 antigen, in the absence of other pan-T-cell
`markers; the B cell is monoclonal with regard to expression of either K or h; and surface
`immunoglobulin (sIg) is of low density. Not only are these characteristics generally adequate for a
`
`CDTL Review,
`Amna Ibrahim M.D.
`NDA # 22249, S-000,
`Treanda (Bendamustine) for CLL
`
`Page 1 of 15
`
`

`

`precise diagnosis, but, importantly, they distinguish CLL from uncommon disorders such as PLL,
`hairy-cell leukemia, mantle-cell lymphoma, and other lymphomasl.
`
`The National Cancer Institute-Sponsored Working Group (NCI—WG) published guidelines for the
`diagnosis and criteria for response for CLLI. The peripheral blood should exhibit an increase in the
`number of small mature-appearing lymphocytes to >5 ,000/pl. The bone marrow aspirate smear must
`show >30% of all nucleated cells to be lymphoid. Although a bone marrow examination is rarely
`required to make the diagnosis of CLL in general practice, it may be evaluable prior to the start of
`treatment in order to define prognostic factors. Subsequently, a bone marrow examination is indicated
`primarily to evaluate response to treatment or to assess normal elements if there is an unexplained
`anemia or thrombocytopenia'. Although not approved for this indication, Fludarabine-based regimens
`are generally used for treatment-naive patients in the US. Chlorambucil and cyclophosphamide are
`approved for this patient population. For further details regarding first-line treatment options, please
`see Medical Officer’s Review (MOR).
`
`2. Background/Regulatory History/Previous Actions/Foreign Regulatory Actions/Status
`
`Bendamustine hydrochloride was developed in the 1960s in former East Germany, but was never
`systematically studied in patients until the 199052. Per applicant, although it has been used for a variety
`of malignancies in Germany for over 30 years, re-approval was required by the German law post
`reunification due to regulatory requirements in the German Democratic Republic. Bendamustine is ‘
`currently marketed in Germany and Bulgaria.
`
`This randomized study was conducted entirely in Europe by Ribosepharm Gran, and per applicant, is
`W Cephalon Inc. acquired the rights and
`using their own statistical analysis plan (SAP), analyzed and submitted the NDA to FDA. This SAP
`and its approach were discussed with the FDA in a preNDA meeting. FDA asked for details of the
`SAP, and did not agree (or disagree) to the design due to insufficient information. In an earlier EOP2
`meeting with Cephalon, FDA agreed to accept a single randomized study and encouraged the use of an
`independent response review committee for efficacy evaluation. For further details regarding the
`regulatory history, please see Dr. Qin Ryan’s Medical Officer’s Review (MOR). There was no Special
`Protocol Assessment (SPA) conducted for the protocol.
`'
`
`3. CMC/Microbiology
`
`The CMC review was completed by Ravindra K. Kasliwal Ph.D., and cosigned by Ravi Harapanhalli
`on 2/27/2008. Their recommendation is as below:
`
`“The application is recommended for an approval action for chemistry, manufacturing and controls
`under section 505 of the Act, provided trademark and labeling acceptability has been determined by
`Office of Drug Safety (DMETS) and provided the manufacturing sites are deemed acceptable for
`cGMP compliance. The product quality microbiology has recommended approval on 06-Feb-2008. The
`recommendation for Office of Compliance regarding the acceptability of the manufacturing facilities is
`pending as of the date of this review.”
`
`
`
`CDTL Review,
`Amna Ibrahim M.D.
`NDA # 22249, S-000,
`Treanda (Bendamustine) for CLL
`
`Page 2 of 15
`
`

`

`3.1 General product quality considerations
`
`According to the microbiology reviewer, Anastasia G. Lolas, MS, and co-signed by Stephen Langille
`Ph.D., this NDA is recommended for approval (Date archived: 2/6/2008) from microbiology point of
`view.
`.
`
`3.2 Facilities review/inspection
`
`According to C. Cruz, thefacilities inspection wasfound acceptable (memo dated March 17“, 2008).
`
`3.30ther notable issues
`
`Per CMC review, the company has not provided data showing compatibility of the constitution
`solution, Sterile Water for Injection, USP, with other commonly available diluents such as -‘""‘
`M .The data (assay and impurity profile) should be provided as part of the
`phase 4 commitment within 6 months of approval of the application (comment for company is provided
`at the end of this review).
`
`It is recommended in the CMC review that the following be included in the action letter:
`
`“We remind you of your agreement in an amendment dated 12-Feb-2008 to initiate change controls for
`all the documents impacted by the revision to the maximum hold time not to- exceed H
`. K and to submit appropriate post-approval correspondence reflecting
`this change.”
`'
`
`4. Nonclinical Pharmacology/Toxicology
`
`Pharmacology/Toxicology Review and Evaluation was signed by Anwar Goheer Ph.D. and cosigned
`by team leader John Leighton Ph.D. on 2/27/2008. According to the review, the nonclinical studies are
`sufficient to support the approval of this NDA. Excerpt from his review states:
`
`“A. Recommendation on approvability: The non-clinical studies submitted to this NDA provide
`sufficient information to support the use of Treanda ® (bendamustine hydrochloride) for the treatment
`of patients with chronic lymphocytic leukemia (CLL).”
`“B. Recommendation for nonclinical studies: No additional non-clinical studies are required.”
`“C. Recommendations on labeling: A separate review will be conducted.”
`
`4.1 General nonclinical pharmacology/toxicology considerations
`
`According to Dr Goheer’s review, “Bendamustine hydrochloride [Treanda®, Cytostasan® (Germany),
`and Ribomustine® (Germany)] belongs to bifimctional nitrogen mustards. Nitrogen mustard and its
`derivatives are alkylating agents which dissociate into electrophilic alkyl groups. These groups form
`covalent bonds with electron-rich nucleophilic moieties. The bifunctional covalent linkage produced
`can lead to cell death via several pathways. The precise mechanism of action of bendamustine has not
`been fully characterized.”
`
`
`
`CDTL Review,
`Amna Ibrahim M.D.
`NDA # 22249, S-000,
`Treanda (Bendamustine) for CLL
`
`Page 3'of 15
`
`

`

`4.2 Carcinogenicigg
`
`As observed in Dr. Goheer’s review, bendamustine is a genotoxic alkylating agent. Oral administration
`for four days induced mammary carcinoma and pulmonary adenomas in mice.”
`
`4.3 Reproductive toxicology
`
`Dr. Goheer’s review also stated that embryo-fetal developmental studies were not conducted by the
`sponsor. During embryo-fetal developmental toxicity study, intraperitoneal administration of
`bendamustine produced embryotoxic and teratogenic effects in mice.
`4.4 Other notable issues
`
`.Per Dr. Goheer’s review, nonclinical safety issues relevant to clinical use were reduction in WBC and
`lymphocytes were observed in a dose related manner during pivotal repeat dose toxicity studies in rats
`and dogs. Treatment related microscopic changes were seen in kidneys (tubular degeneration/necrosis)
`in both species. Cardiomyopathy (focal/multifocal) was observed in male rats only. Heart rates of dogs
`at 6.6 mg/kg/day were reduced during cycle 2 (2 males & 1 female, 3/6 animals). A vigilant monitoring
`of QT prolongation is warranted until more clinical experience is gained. Bendamustine is mutagenic,
`carcinogenic, and teratogenic like other nitrogen mustard alkylating agents. There are no outstanding
`issues noted in Dr Goheer’s review.
`
`5 Clinical Pharmacology/Biopharmaceutics
`
`Julie Bullock, Ph.D. was the clinical pharmacology reviewer for this NDA. Her recommendations were
`as follows:
`
`“This NDA is considered to be deficient from a clinical pharmacology perspective due to the lack of
`data available regarding pharmacokinetics at the proposed dose, dose proportionality, human excretion
`and metabolism, effect on QT prolongation, in-vivo drug-drug interactions, and in-Vitro p—glycoprotein
`screens.” She recommended that the NDA will be considered acceptable pending the sponsor’s
`agreement to four Phase 4 commitments. “No pharmacokinetic data was obtained at the proposed dose
`(100 mg/m2 IV over 30—mins) in the proposed CLL patient population. In addition, there were no
`formal PK dose ranging studies, and no multipledose pharmacokinetic assessments. A mass-balance
`study in humans was initiated in 2008. Completion of the mass balance study, assessment of QT
`prolongation, ' m ', in—vitro p-gp substrate and inhibition screens, and
`in-vivo interaction studies with a CYP1A2 inhibitor and inducer will be phase 4 commitments. The
`completion of renal and/or hepatic studies will depend on the outcome of the mass balance evaluation.” I
`Dr. Bullock’s review was cosigned by Brian Booth Ph.D. on 2/22/2008.
`
`5.1 General clinical phannacology/biopharmaceutics considerations:
`
`The Clinical pharmacology assessments were mostly based on Studies conducted in patients with Non-
`Hodgkin’s Lymphoma (NHL). As noted by Dr. Booth, Team Leader/Deputy Director, Clinical
`Pharmacology, “the dose in this patient population is 20% higher than the dose used for CLL patients
`(100 mg/mz). In the NHL patients, the higher dose was infused for 1 hour (twice as long as CLL
`patients) and had an elimination half-life of about 5 hours. The Cmax was approximately 5600 ng/ml,
`and the AUC was approximately 6600 ng-hr/ml. No information is available from these studies
`regarding the dose proportionality of bendamustine. Bendamustine generates two active metabolites,
`named M3 and M4. However, concentrations of these metabolites in vivo appear to be 1/10 (M3) and
`
`
`
`CDTL Review,
`Amna Ibrahim M.D.
`NDA # 22249, S-000,
`Treanda (Bendamustine) for CLL
`
`Page 4 of 15
`
`

`

`1/ 100th (M4) of the concentrations of the parent, and are not likely to contribute significantly to the
`activity of the drug.”
`
`5.2 Drug-drug interactions
`
`Per Dr Booth, “Based on in vitro studies, cytochrome P-450 1A2 (CYP 1A2) appears to mediate the
`metabolism of bendamustine. No in vivo drug-drug interactions were conducted to assess the impact of
`co-administen'ng medications that induce CYP 1A2 (e.g. smoking, omeprazole) which could reduce
`bendamustine plasma concentrations to sub-therapeutic levels. Also, no in vivo studies were conducted
`to assess the effect of CYP 1A2 inhibitors (e. g. fluvoxamine, ciprofloxacin), which might be expected
`to cause toxicity via increased plasma levels of bendamustine. The role of P-glycoprotein and other
`transporters in the fate of bendamustine was not assessed.”
`
`5.3 Pathway of Elimination
`
`No “mass balance” study was conducted, and the routes of elimination and the extent that these routes
`play in the elimination of the drug were not determined. Therefore, although metabolism plays some
`role in the elimination of the drug, extent of renal elimination remains unknown. Therefore, the need
`for studies to assess the effect of organ dysfunction on drug elimination has not yet been assessed.”
`
`5.4 Demographic interactions/special populations
`
`According to Dr. Booth, a pharmacokinetic study of bendamustine in Japanese patients was also
`undertaken. This study revealed a similar disposition of bendamustine, but Japanese patients had a
`mean clearance that was 20% lower than the North American population. The clinical significance of
`this finding is unclear.
`
`5.5 Thorough QT study or other QT assessment
`
`QT assessment has not been conducted and will be a post-marketing commitment
`
`5.6 Other notable issues
`
`See post-marketing commitments.
`
`6 Clinical Microbiology
`
`Not Applicable.
`
`7 Clinical/Statistical
`
`7.1 Efficacy
`
`7.1.1 Dose identification/selection and limitations
`
`Per applicant, two dose finding studies were conducted in previously treated patients with refiactory or
`progressive Binet stage B or C CLL with 15 and 16 patients respectively. In one study in which
`bendamustine was administered every 21 days, MTD was 70 mg/mZ/day administered on day l and 2.
`Based on these results, it was recommended that the treatment cycle be prolonged to 28 days. In the
`Page 5 of 15
`
`CDTL Review,
`Amna Ibrahim M.D.
`NDA # 22249, S-OOO,
`Treanda (Bendamustine) for CLL
`
`

`

`second study with younger and less heavily treated patients, the MTD was 100 mg/m2 on day 1 and 2,
`when given every 21 days. A recommendation was made in this study as well, to prolong the treatment
`cycle to 28. Later, in a Scientific Protocol Review Board Meeting, experts recommended that 100
`mg/m2 be administered on days 1 and 2 and repeated every 28 days.
`
`7.1.2 Randomized, Phase 3 clinical study
`
`The applicant submitted a single randomized study titled “Title ofStudy: Phase III, Open—Label,
`Randomized, Multicenter Efficacy and Safety Study ofBendamustine Hydrochloride Versus
`Chlorambucil in Treatment—Nai've Patients With (Binet Stage B/C) B-CLL Requiring Therapy”. This
`study was conducted from S‘h November 2002 through 26th March 2006. The safety and efficacy of
`bendamustine were evaluated in trial comparing bendamustine to chlorambucil. The trial was
`conducted in 301 previously-untreated patients with Binet Stage B or C (Rai Stages I - IV) CLL
`requiring treatment. Need-to-treat criteria included hematopoietic insufficiency, B-symptoms, rapidly
`progressive disease or risk of complications from bulky lymphadenopathy. Patients with autoimmune
`hemolytic anemia or autoimmune thrombocytopenia, Richter’s syndrome, or transformation to
`prolymphocytic leukemia were excluded from the study.
`
`Table: Dose and schedule of bendamustine and chlorambucil
`Treatment Arms
`Dose and Schedule
`
`{1:35:21
`
`100 mg/m2 on days 1 and 2 during each 28-day cycle.
`
`N= 1 48
`
`Chlorambucil
`
`0.8 mg/kg (Broca’s normal weight) orally on days 1 and 15 or, if
`necessary, divided doses on days 1 and 2 and on days 15 and 16 during
`each 28-da c cle.
`
`Co-primag endpoints:
`- Overall response rate (ORR) based on the criteria as defined by the NCI—WG on CLL
`(original protocol)
`Progression-free survival (PFS)
`
`-
`
`The primary efficacy endpoints were overall response rate (OR) and progression-free survival (PFS)
`assessed for the intent-to-treat (ITT) population. Originally, these were based on the NCI—WG criteria
`for CLL. In amendment #4, the primary endpoints were to be based on ICRA assessment. Overall
`response rate (OR) was defined as the proportion of patients in each treatment group with a best
`response of complete response (CR), nodular partial response (nPR), or partial response (PR) to
`treatment. Progression-free survival (PFS) was defined as the time from randomization to progressive
`disease (PD) or death for any cause, whichever occurred first.
`
`Analysis of the co-primary endpoints will be described using three different major analyses, i.e., as
`assessed by Independent Response Assessment Committee (ICRA), investigator’s assessment and
`“calculated assessment”. The “calculated assessment” was the applicant’s efficacy assessment that used
`a computer programmed algorithm based on NCI-WG Criteria for CLL with elements obtained from the
`source documents.
`
`CDTL Review,
`Amna Ibrahim M.D.
`NDA # 22249, S-000,
`Treanda (Bendamustine) for CLL
`
`Page 6 of 15
`
`

`

`Secondm endpointsszwere
`time to progression (TTP)
`duration of response
`overall survival (OS)
`infection rate
`
`-
`-
`-
`
`-
`-
`
`quality of life
`toxicities
`
`The secondary endpoints will not be discussed further in this review. Please see MOR for details.
`
`Results:
`
`The demographics were unremarkable except that all but one patient were Caucasians. Seventy percent
`had Binet Stage B disease and 30% had stage C disease. Ninety percent patients had co-expression of
`CD5, CD23 and either CD19 or CD20 or both. Two of the largest enrolling centers were found by the
`applicant to have several major deviations from the protocol. These sites enrolled 54 patients.
`
`Efficacy
`Applicant’s table
`
`PrimaryEfficacy
`Variable
`
`TREANDA
`1 =153
`
`Chlorambucil
`. '=148
`
`pwalue
`
`Resp onse Rate by Independent Refiéw, n (%)
`Overall response rate
`95 (62)
`(95%CI)
`(54.40, 69.78)
`Complete response (CR)
`42 (27)
`Nodular partial response (nPR)
`15 (10)
`Partial response (PR)
`38 (25)
`Resp onse Rate by Calculated Analys‘s, n (%)
`Overall response rate
`90 (59)
`(95%CI)
`(51.03= 66-62)
`Complete response (CR)*
`13 (8)
`Nodular partial response (nPR)
`4 (3)
`Partial res case
`73 48
`
`.
`
`49 (33)
`(25.53. 40.69)
`3 (2)
`4 (3)
`42 (28)
`
`38 (26)
`(13.64= 32.71)
`1 (<1)
`0
`37 2)
`
`Progression-F ree Survival by Independent Review
`9
`Median, months
`21
`Hazard ratio (95% CI)
`0.23 (0.13 - 039)
`Progression—F tee Surviml by Calculated Mai:
`6
`Median, months
`18
`. Hazard ratio (95" n C
`027 (0.17 — 0.43)
`CImfidmce interval
`*CR onlyarsignad inpatients withraquisita boneW sample firmimtion.
`
`<0.0001
`
`<0.0001
`
`,
`
`<0.0001
`
`<0.0001
`
`CDTL Review,
`Amna Ibrahim M.D.
`NDA # 22249, S-000,
`Treanda (Bendamustine) for CLL
`
`Page 7 of 15
`
`

`

`Table: Efficacy Data based on Calculated Algorithm based on NCI—WG Criteria for CLL
`FDA Statistical Reviewer’s table
`
`
`
`‘
`
`Res onse Rate 11 %
`
`=153
`
`=14s
`
`P
`
`
`
`
`
`-—l—I_I_
`
`
`m—
`m“—
`
`
`m-
`
`
`Pro ; ression—Free Survival
`6 5.6, 8.6
`18 11.7, 23.5
`Median, months 95% CI
`
`
`<0.0001
`0.27 0.17, 0.43
`Hazard ratio 95% CI
`
`*CR was defined as peripheral lymphocyte count _<_ 4.0 x 109/L, neutrophils 3 1.5 x lOg/L, platelets >100 x 109/L,
`hemoglobin > 110g/L, , absence of palpable hepatosplenomegaly, lymph nodes _<_ 1.5 cm, < 30% lymphocytes
`without nodularity in at least a normocellular bone marrow and absence of “B” symptoms
`
`Response Rate (RR):
`
`The overall response rate was approximately 60% in the bendamustine arm and 26% -33% in the
`chlorambucil arm as assessed by ICRA or by the calculated algorithm based on NCI—WG CLL criteria
`respectively (see tables above). The CRs were 27% according to ICRA assessment and 8% according
`to the calculated algorithm. For RR according to investigator’s assessment, please see MOR. The
`clinical and statistical teams recommend that the efficacy findings be based on the calculated algorithm
`that used NCI-WG criteria for CLL for the labeling for reasons discussed below.
`
`ICRA assessed versus calculated algorithm-based analysis:
`
`According to applicant, the original sponsor supplied the ICRA with the tumor evaluations from all
`patients blinded for patient name, center, treatment arm and overall response assessment. Each ICRA
`member evaluated all patients separately. The assessments were provided to the sponsor who compared
`the assessment of the ICRA members. Patients with identical assessments were to have analysis entered
`directly. Responses assessed differently underwent a discussion and consensus process between the
`ICRA members. The analysis with the consolidated overall response was used. However, no records on
`the decision making were kept, and therefore, the results based on the ICRA assessment can not be
`verified.
`
`'/": [he R by this assessment has
`e the ICRA-assessed analyses
`"J
`a markedly higher number of CR5, although the OR is similar to that obtained by the calculated
`algorithm. However, the results according to ICRA assessment are not verifiable
`‘
`\
`
`
`CDTL Review,
`Amna Ibrahim M.D.
`NDA # 22249, S-000,
`Treanda (Bendamustine) for CLL
`
`Page 8 of 15
`
`

`

`/
`
`/
`
`
`,. ICRA did not adhere to the
`.51:
`_
`-
`"T 7’
`L
`'
`NCI—WG criteria as specified in the original protocOl. The reasons for deviation from NCI—WG criteria
`for individual patients were not captured. The results based on ICRA assessment are not verifiable.
`
`Progression-free Survival:
`
`The median PFS based on the ICRA assessment was 21 months in the bendamustine arm and 9 months
`
`in the chlorambucil treatment group; the difference between treatment groups in PFS was statistically
`significant in favor of bendamustine treatment. Hazard ratio for this difference was 0.23. For the NCI-
`WG criteria based calculated analysis, the median PFS was 18 months on the Bendamustine arm and 6
`months on the chlorambucil arm with a hazard ratio of 0.27. The difference in median PFS was 12
`
`months between the two groups in both analyses.
`
`Figure: PFS based on prespecified algorithm based on NCI-WG criteria
`
` SurvivalDistributionFunction
`
`
`
`15
`
`20
`
`25
`
`30
`
`35
`
`40
`
`45
`
`
`
`0
`
`5
`
`10
`
`Progression—Free Survival (months)
`
`SuioyTreaimem '"TREANDA
`
`“Chlorambucil
`
`7.1.3 Discussion of primary and seconm reviewers’ comments and conclusions
`
`The primary clinical efficacy reviewer Dr. Qin Ryan and clinical safety reviewer Ms. Virginia
`Kwitkowski M.S.,, R.N., C.R.N.P., both recommend approval of bendamustine for CLL pending
`completion of financial disclosure by the applicant.
`
`
`
`‘
`CDTL Review,
`Amna Ibrahim M.D.
`NDA # 22249, S-OOO,
`Treanda (Bendamustine) for CLL
`
`Page 9 of 15
`
`

`

`The efficacy analyses of the primary endpoint were statistically significant in favor of bendamustine
`and these included the following
`- RR and PFS based on prespecified NCI-WG criteria,
`- RR and PFS based on ICRA,
`- RR and PFS based on analyses excluding of the 2 sites with major violations.
`
`As noted by Ms Kwitkowski, the requirements for blood transfusions (20% on Treanda arm), decreased
`with increasing number of treatments. This improvement in anemia was most likely due to disease
`response. Patients with Patients with complete responses (CR5) had more improvement in hemoglobin
`than those with partial responses (PRs).
`
`Overall survival was immature at the time of data cut-off.
`
`7.1.4 Pediatric use/PREA waivers/deferrals
`
`CLL is a disease that generally affects older individuals. A pediatric waiver was given to Treanda for
`this indication.
`
`7.1.5 Discussion of notable efficacy issues
`
`The efficacy of bendamustine has been demonstrated by a clinically and statistically significant
`improvement in response rate and progression-free survival. The response rate improved to 59% with
`an 8% CR rate. A twelve month improvement in median progression-free survival was observed. No
`drug has demonstrated an unequivocal improvement in overall survival to date.
`
`7.2 Safety
`
`7.2.1 General safeg considerations
`
`The safety population included 153 patients treated on the Treanda arm. The population was 45-77
`years of age, 63% male, 100% white, and had treatment naive CLL. Adverse reactions were reported
`according to NCI CTC v.2.0.
`
`According to the Safety reviewer Virginia Kwitkowski, “Non-hematologic adverse reactions were
`mostly of low grade (1-2). Eighty-eight (58%) patients in the bendamustine treatment group and 44
`(31%) patients in the chlorambucil treatment group reported at least one grade 3 or 4 adverse reaction.
`Both grade 3 and 4 adverse reactions occurred more frequently in the bendamustine treatment group
`than in the chlorambucil treatment group. Grade 3 events were reported in 33% of the bendamustine
`patients as compared to 22% in the chlorambucil patients. Grade 4 events were reported in 25% of
`patients in the bendamustine group as compared to 8% of patients in the chlorambucil group”
`
`7.2.2 Safegg findings from submitted clinical trials.
`
`In the randomized CLL clinical study, hematologic adverse reactions (any grade) in the Treanda group
`that occurred with a frequency greater than 15% were neutropenia (28%), thrombocytopenia (23%),
`anemia (19%), and leukopenia (18%). Non-hematologic adverse reactions (any grade) in the Treanda
`group that occurred with a frequency greater than 15% were pyrexia (24%), nausea (20%), and
`vomiting (16%). Grade 3 or greater hematology laboratory test abnormalities were anemia (13%),
`
`Page 10 of 15 ‘
`
`CDTL Review,
`Amna Ibrahim M.D.
`NDA # 22249, S-OOO,
`Treanda (Bendamustine) for CLL
`
`

`

`thrombocytopenia (11%), and decreased neutrophils (43%). The incidence of febrile neutropenia was
`6% and 20% of the patients were transfused with RBCs. The most frequent adverse reactions leading to
`study withdrawal for patients receiving Treanda were hypersensitivity (2%) and pyrexia (1%).
`
`“Grade 3/ 4 hematologic adverse reactions with a frequency greater than 10% in the bendamustine
`treatment group were neutropenia (24%), leukopenia (15%), and thrombocytopenia (13%).
`Grade 3/4 non-hematologic adverse reactions were reported by 52 (34%) patients in the bendamustine
`treatment group and 25 (17%) patients in the chlorambucil treatment group. Grade 3/ 4 non-
`hematologic adverse reactions with a frequency greater than 1% in the bendamustine treatment group
`were pyrexia (4%), pneumonia (3%), rash (3%), hypertension (3%), hypertensive crisis (2%),
`hyperuricemia (2%), and infection (2%). Five patients (3%) in the bendamustine treatment group
`experienced febrile neutropenia compared with none in the chlorambucil group. Neutropenic infection
`occurred in 10 bendamustine patients compared with l in the chlorambucil group. There were 2 events
`of grade 3 sepsis, both in patients in the bendamustine treatment group.'Both patients recovered. Grade
`3/4 hematologic adverse reactions with a frequency greater than 5% in the chlorambucil treatment
`group were neutropenia (9%) and thrombocytopenia (8%)”.
`
`Myelosuppression, infections related to myelosuppression, tumor lysis syndrome, hypersensitivity
`reactions, hypertension and other cardiac events and secondary malignancies were identified as
`significant adverse reactions in the safety review.
`
`“Thirty-four deaths occurred during the conduct of study 02CLLIII. An equal number (17) of deaths
`occurred in each treatment group” and “The most common attribution for death was progression of
`disease (41% of patients in each group). Four patients died during the treatment phase of the study or
`within 30 days of the last study drug dose, one patient in the bendamustine group and three patients in
`the chlorambucil group.”
`
`According to Ms Kwitkowski, “Twenty-two patients were withdrawn from the study because of
`adverse reactions; 17 (11%) patients who received bendamustine and 5 (3%) patients who received
`chlorambucil. The most frequent adverse reactions causing withdrawal were hypersensitivity
`(occurring in 3 bendamustine patients and 1 chlorambucil patient) and pyrexia (occurring in 2
`bendamustine patients and 1 chlorambucil patient .”
`
`She also cements: “Non-clinical studies described dose-related cardiac toxicity in animals. The
`_
`clinical dose-escalation studies demonstrated dose-related cardiac toxicities. Though the overall
`incidence of cardiac toxicity in the bendamustine arm was low (and similar to the chlorambucil arm);
`bendamustine may have cardiac toxicities, particularly at higher doses than those utilized in the pivotal
`trials for CLL and NHL. The large variety of cardiac events reported in these smaller studies make it
`difficult to provide firm attribution to bendamustine. M ,
`
`ECG monitoring in this study was not adequate to
`evaluate the potential for QT prolongation because ECGs were only obtained at baseline and end of
`study; and interval measurements were not obtained.”
`
`In the amended Medical Oflicers ’ Review, Ms. Kwitkowski observed “In the dataset exploration, no
`cases that met Hy ’s law criteria were found. These results do not exclude the potentialfor DILI due to
`the small sample size, but no evidencefor DILI was identified during this data exploration. "
`
`CDTL Review,
`Amna Ibrahim M.D.
`NDA # 22249, S-OOO,
`Treanda (Bendamustine) for CLL
`
`Page 11 of 15
`
`

`

`7.2.3
`
`Safety update
`
`Per Ms Kwitkowski: “The postrnarketing data provided by the Applicant does not provide new safety
`concerns that would affect the regulatory decision to approve bendamustine for CL .”
`
`7.2.4 Discussion of primag reviewer’s comments and conclusions
`
`I concur with the primary reviewers’ conclusions.
`
`7.2.5
`
`Pre-Approval Safety Conference
`
`A Pre-Approval Safety Conference was conducted on 2/27/2008. The findings of the safety reviewer
`were discussed and are included in this review above and in Ms. Kwitkowski’s review of safety.
`
`7.2.5 Discussion of notable safefl issues
`
`See section 7.2.1.
`
`Statistics Review Team’s comments:
`
`According to Shenghui Tang, Ph.D., primary statistics reviewer, “A total of 302 patients were screened
`and 301 were randomly assigned to treatment (1 patient was not assigned to a treatment group due to
`refusal) at 45 centers throughout 8 countries. The sponsor reported that the proportion of patients with
`OR was 62% in the bendamustine treatment group compared with 33% in the chlorambucil treatment
`group (p<0.0001) as determined by the Independent Committee for Response Assessment (ICRA). The
`primary PFS analysis showed that the bendamustine treatment was superior to chlorambucil treatment
`(median 21 vs. 9 months, hazard ratio (HR) 0.23, p<0.0001). Based on the data submitted by the
`sponsor these results were confnmed by this reviewer and the data support the efficacy claim.”
`
`“Whether the endpoints and the sizes of the effects on these two endpoints in this phase III study are
`adequate for approval is a clinical decision.” (Review dated 2/ l 9/2008, concurrence signatures
`provided by Dr. Sridhara and Dr. Chakravarty on the same dates).
`
`According to Raj eshwari Sridhara Ph.D., Team Leader and Deputy Director for Biometrics, “I concur
`with the primary reviewer, Dr. Tang’s conclusion that the data submitted supports the claim that
`bendamustine has demonstrated superior overall response rate (OR) and progression-free survival
`(PFS) compared to chlorambucil (OR of 59% vs. 26% and PFS HR = 0.52, p-value < 0.0001).”
`
`“Progression-free survival was assessed by a panel of three independent expert hematologic oncologists
`and also objectively calculated using an algorithm based on NCI working group criteria. According to
`the sponsor, in performing the review the members of the independent panel were allowed to exercise
`clinical judgment in determining response.” and “The FDA reviewers were able to verify the calculated
`response rates and PFS, but could not verify the same as determined by the independent panel due the
`subjective nature of the independent evaluation. Therefore, it is recommended that the calculated
`response rates and PFS estima