clinical adverse events, by body system and COSTART terminology. If a patient
`experienced more thanl episode of an adverse event, the patient was counted only once
`for that event. If a patient had more than 1 adverse event in a body system category, the
`patient was counted only once in that body system total.
`
`29
`
`APPEARS THlS WAY
`
`0N ORIGINAL
`
`

`

`30
`
`
`
`No. (%) of Patients with Clinical Adverse Events by Body
`S stern: ALA-O18
`
`
`Body System Catcgory/
`W W (coswm
`———
`——-Elm-
`—_—
`
`LEVULAN®
`
`Vehicle
`
`__-l%-
`_--nm-
`——-nm-
`m—-—
`_--1m-
`——-ma-
`-}_——
`_—-Im-
`“——
`
`
`
`_——
`
`«_—
`———
`
`)
`_—
`10%)
`
`0(
`
`Hemic and Lymphatic System
`Abnormal Platelets
`
`Metabolic and Nutritional System
`
`Musculoskclctal System
`
`—
`
`10%)
`10%) —
`0(
`_
`
`10%)
`10%)
`
`.
`
`10%) '
`10%)
`
`—
`
`0(
`
`
`
`
`
`
`
`
`
`'
`
`
`
`
`
`APPEARS THIS WAY
`0N ORlGlNAL
`
`

`

`No. (%) of Patients with Clinical Adverse Events by Body
`8 stem: ALA-O18 (Continued)
`Body System Category/
`
`LEVULAN®
`
`31
`
`Vehicle
`
`
`
`
`
`———
`3 (10%)M
`
`—_—-ma-
`——-I%-
`——-I%_
`
`
`
`_—
`specter Senses
`o (0%)
`2 (2%)
`0 (0%)
`1 (1%)
`Cataract NOS
`0 (0%)
`Conjunctivitis
`1 (1%)
`‘ Unevaluablc event: This patient had 3 procedures associated with surgery (Pt.18219,
`Table 16.2.13)
`
`The five patients in the LEVULAN® arm who experienced serious adverse events
`(patients 18102, 18219, 18221, 18402, and 18501) experienced broken lefi leg (from
`accident), implant of a thalamic stimulator (to treat a tremor), a pre-existing
`hyperkeratotic (Grade 3)actinic keratosis (at an untreated site), pre-existing squamous cell
`carcinoma on the left ear, and ruptured abdominal hernia. Invesfigators considered these
`episodes unrelated to exposure to LEVULAN®.
`
`Conjunctivitis developed in one patient (no. 18711) 44 days afier LEVULAN®
`application. The investigator concluded that it was not related to LEVULAN® treahnent.
`Reviewer’s Comment:
`
`Reviewer it Comment:
`With most topical medications, there is only one relevant period during which adverse
`events need to be assessed:
`the period ofdrug administration. In contrast, with
`LEVULAN®, three distinctperiods must be assessedfor the incidence and severity of
`adverse events: (a) [pre-PDI] the periodfrom LEVULAN® application until light
`administration; (b) [peri-PDTIthe period during and shortly after light administration;
`and (c) [post-PDUtheperiodfrom shortly after administration oflight therapy to end-of-
`foIIow-up. The reason why adverse events within each ofthese three time periods must be
`
`
`
`lim—
`
`M«
`
`_—
`-E_——
`_—_
`
`
`
`
`
`
`'
`
`

`

`32
`
`assessed separately is that the type and quantity ofthe adverse events differ. A timepoint
`usefulfor separating short-fiom long-term adverse events would be 24 hours after
`treatment, becausefor most patients the pert-PDT adverse events have resolved or
`returned to baseline by this timepoint.
`Another complication in assessment ofadverse event incidence and severity is that
`untreated actinic keratoses manifest some ofthe signs (e.g. erythema, hyper— and
`hypopigmentation) that are considered adverse events.
`‘
`Adverse Events: pre-PDT
`
`The percentage ofpatients who develop signs and/or symptoms of a photodynamic
`response (e.g., buming/stinging and/or edema) during the time interval between
`application of LEVULAN® and administration ofblue light may be an indirect measure
`of the prevalence of an (inappropriate) photodynamic response that results from
`inadvertent exposure to ambient light. Despite the presence of information in the
`protocol that wams patients to protect the lesions being treated fi'om light exposure for a
`minimum of 14 to 18 hours after application (i.e., “to avoid direct exposure of target sites
`to sunlight or other high intensity light sources, including tanning light devices”), active-
`treated patients do manifest signs/symptoms of a photodynamic response prior to blue
`light administration: 47% of active treatment patients develop burning/stinging between
`Baseline A and B, while 14% of vehicle treatment patients develop buminystinging;
`17% of active treatment patients develop edema between Baseline A and B, while no
`vehicle treatment patients develop edema between Baseline A and B.
`
`Reviewer '5' Comment: The incidence and/or increasedprevalence ofburning/stinging
`and edema that develops between Baseline A and B is attributable either to irritancy or
`to an inappropriate photodynamic response. Since no irn'tancy study has been performed
`with the to-be-marketedfonnulation, it is not possible to discern which ofthese two
`alternative explanations is correct.
`
`Adverse Events:' periiPDT ‘*
`
`-
`
`'
`
`-
`
`-— - ~ 7
`
`»
`
`The medical reviewer analyzed the incidence of signs and/or symptoms expected during a
`photodynamic response fiom the period including Baseline B until 24 hours after light
`treatment for patients undergoing active and control treatment, as is depicted below.
`
`APPEARS nus WA»!
`on ORIGINAL
`
`

`

`BEST POSSIBLE’COPY
`
`as
`
`
`
`
`
`Edema?
`
`ALL
`
`(11%)
`23/72
`
`3/72
`
`(88%)
`12/72
`\ V
`68/72
`
`(57%)
`0/21
`
`7/21
`
`(33%)
`0/21
`
`2/21
`
`(19%)
`5/16
`
`0/16
`
`(81%)
`3/16
`
`16/16
`
`(50%)
`0/8
`
`1/8
`
`
`
`
`
`
`(50%)
`0/8
`
`2/8
`
`.ncidence of erythema, edema, and burning/stinging during and/or 24 hours after photodynamic
`therapy‘: ALA-018
`
`
`
`
`Fraction of
`
`patients with some
`
`
`or all target lesions
`
`involved:
`
`
`--- - 61%) (19%)--
`
`--------
`' Burning/Stinging
`
`
`
`
`*defined as the prevalence of adverse events during the time points at baseline B, through light treatment, and
`at 24 hours after light treatment
`
`
`
`
`fSponsor has not collected data that would permit the classification ofthese adverse events as mild, moderate,
`or severe.
`
`
`From Data Listings 15, 16, and 18, Vol. 1.56
`
`Sponsor’s analysis of the data listings (not shown) largely corroborated the medical
`officer’s analysis. The consequence of treatment with LBVULAN® and blue light was to
`increase the prevalence ofpatients who experienced erythema, edema, and
`stinging/burning associated with treated lesions in the period during and shortly after
`photodynamic therapy.
`_ -.
`,
`7 _
`
`As depicted in the following table, apprdximately half of the patients in whom all the
`target lesions were erythematous during and/or shortly after photodynamic therapy did
`not have all their target lesions erythematous atone week after treatment. The adverse
`events of edema and burning/stinging resolved more quickly, usually within 24 hours
`after completion of light therapy.
`
`APPEARS 1111s WAY
`on 0111mm
`
`

`

`BEST POSSIBLE COPY
`
`34
`
`Patients with adverse events involving 100% of target lesions during photodynamic therapy:
`prevalence of erythema, edema, and burning/stinging at 24 hours and one week after therapy:
`ALA-O18
`
`
`
`
`
`
`
`—.Eli'l_ VEHICLE
`
`Fraction of
`24 hours
`One
`24 hours
`One
`
`patients with all
`week
`..
`week
`target lesions
`
`involved:
`
`V
`
`
`ACTIVE
`VEHICLE
`24 hours
`One
`24 hours
`One
`week
`week
`
`
`
`
`
`
`
`
`Erythema
`
`30/64
`(47%)
`———
`Burning/Stinging
`17/68
`1/68
`
`7/14
`2/7
`(50%)
`(29%)
`_ - _
`1/2
`
`2/4
`(50%)
`
`
`
`
`(1%)
`(25%)
`From Data Listings 15, 16, and 18, Vol. 1.5
`
`6
`
`___,,,,, *
`
`In response to medical reviewer’s request, sponsor submitted clinical slides showing the
`time course of healing in some of the patients treated with LEVULAN® in this trial. As
`judged from these slides, in several of these patients the erythema and edema arising
`' riposfiPDT appearedto efifidbeyondthe borders of the actinic keratosis. Of note, the
`application instructions in the protocol state: “gently dab the/lesions designated to be
`treated being certain to uniformly .wet the entire lesion including the margin (Appendix E:
`Application Instructions). The medical reviewer recognizes that it is difficult to assess
`fiom clinical slides the marginsofthe actinic keratoses, so it is possible that the
`erythematous and edernatous reaction actually was confined to the areas treated with
`LEVULAN®, but the reviewer thinksjtmore likely that the reaction can extend beyond
`the borders. In some patients (#18103, #18302), the area with erythema and edema
`appears two- to three— fold larger than the area with the actinic keratosis.
`
`./
`
`'
`
`The following table depicts the degree of severity ofbuming/stinging during and/or 24
`hours after photodynamic therapy. The majority of treated patients characterized as
`severe the degree of burning/stinging on at least one target lesion during treatment.
`
`APPEARS THIS WAY
`
`0N ORlGlNAl.
`
`

`

`0/8
`‘
`
`
`
`
`
`
`
`
`
`asst POSSIBLE copy
`
`35
`
`0/21
`
`
`
`——
`
`Severity:
`.
`Burning/Stinging
`
`severity of Burnino-IStining durin- and/or 24 hours after photodynamic therapy: ALA-018’ *
`
`
`
`
`
`— VEHICLE _mez=-
`MILD/
`SEVERE
`MILD/
`SEVERE
`MILD/
`SEVERE
`MlLD/
`SEVERE
`MODER-
`MODER-
`MODER—
`ATE
`ATE
`ATE
`ATE
`8/16
`8/16
`41/72
`30/72
`10/21
`3/8
`(50%)
`(50%)
`(57%)
`(42%)
`(48%)
`(38%)
`"the fraction ofpatients who experienced burning/stinging on at least one target lesion up to (and not
`exceeding) the degree of severity indicated, during the time period between baseline and 24 hours after light
`
`
`
`
`
`
`
`From Data Listing 17, Vol. 1.56
`
`'
`
`‘
`
`The mean lesion number for patients who experienced severe discomfort was 6.48, while
`the mean lesion number for lesions in the active treatment arm was 6.98. Sponsor did not
`collect information that would permit the reviewer to determine if the severity of
`burning/stinging was related to the precise location within the scalp or face. Some
`patients at all the study centers reported severe discomfort. Patients with exclusively
`Grade 1 lesions, and patients with exclusively Grade 2 lesions, reported severe
`discomfort. Sponsor claims that discomfort [burning/stinging] starts declining in severity
`’during or immediately after light treatment (from label’s information to patients: “these
`feelings of discomfort [photodynamic response], if any, will improve at the end of light
`treatment”). There were 108 treatments (including first treatments and re-treatments) in
`which any burning/stinging was experienced. Following 64 of these occasions,
`burning/stinging was less severe at one minute after treatment compared to the most
`severe pain experienced during treatment. In comparison, on 44 occasions discomfort
`was as severe or more severe at one minute after treatment compared to the most severe
`discomfort experienced during treatment. In a significantly higher proportion of
`' treatments (p<.05, one-sided z-approximation of the binomial distribution), the severity
`of discomfort was less at one minute after treatment compared to the most severe
`discomfort experienced during treatment. No or minimal discomfort was noted by
`patients by 24 hours after treatment.
`
`Sponsor claims that “in the Phase III studies, stinging/burning and discomfort was less
`severe in those patients whose AK lesions were retreated with LBVULAN afier eight
`weeks.” In comparing burning/stinging between the first and second treatments in the
`same patients, the most striking observation is that there are fewer lesions undergoing
`retreatrnent than had undergone first treatment:
`the average number of lesions
`undergoing first treatment is 6.08, with standard deviation of 3.13, while the average
`number of lesions undergoing retreatrnent is 3.25, with standard deviation of 2.78. There
`were five patients who experienced retreatrnent who had the same number oftarget
`lesions at enroriment and at week 8. The sum of the burning/stinging scores during and
`immediately after treatment was calculated for these five patients: 3 had increased scores
`upon retreatrnent, 1 had a decreased score, and 1 had the same score. The major reason
`
`

`

`BEST POSSIBLE COPY »
`
`36
`
`why patients experienced less burning/stinging upon retreatment appears to be that
`patients had fewer lesions treated at Week 8.
`
`Degree of Severity;
`
`Scaling/Crusting
`
`1 (1%)
`
`1 (6%)
`
`1 (6%) “In“
`5 (31%)
`3 (19%)
`2 (25%)
`0
`
`Incidence and Severit of Cutaneous Adverse Events, post-PDT‘“: ALA-018
`FACE
`ACTIVEig
`VEHICLE
`IVEHICL
`MILD/
`SEVERE
`MILD]
`SEVERE
`MILD/
`SEVERE
`MILD/
`SEVERE
`MODER-
`M DER-
`MODER-
`MODER-
`ATE
`ATE
`ATE
`ATE
`49/72
`8 (38%)
`9 (56%)
`3 (38%)
`(68%)
`a
`Pain/Tenderness I. 1 (1°/ .-
`Itching
`27/72
`2 (3%)
`4 (19%)
`(38%)
`4/72
`
`Ulceration
`
`Bleeding]
`Hemorrhage
`Hypo/
`
`:1
`
`6.
`
`(6%)
`2/72
`
`(3%)
`
`1 (6%)
`
`18/72 (25%)
`
`4/21 (19%)
`
`5/16 (31%)
`
`2/8 (25%)
`
`L1i
`
`1
`
`-n
`H"-
`_ (1%) fl-“
`- (3%) "n“
`—--fln-flu
`Scabbing
`3/72
`1 (1%)
`_ (11%,
`1. "-
`Erosion
`8/72
`1 (1%)
`— (3%) ------
`Wheal/Flare
`2/72
`
`Skin disorder, NOS
`
`5/72
`
`l (1%)
`
`1 (5%)
`
`l (5%)
`
`find A‘1;é
`
`1 (13%)
`
`-
`n
`-
`
`(7%)
`*"defined as adverse events that are not present at baseline A. If a patient had an event recorded at more than
`one visit, that patient is counted only once.
`{H‘his category refers to the fraction of patients who deVelop hypo- and/or hyper- pigmentation on at least one
`target lesion during the treatment course. Sponsor has not collected data that would permit the classification of
`‘he hypo- and/or hyper— pigmentation as mild, moderate, or severe.
`
`om Data Listing l3, 19, Vol. 1.56
`
`

`

`37
`
`In the process of generating the above table, reviewer has clustered together related PDT
`responses [crusting, scaling, and hyperkeratosis grouped into scaling/crusting;
`hemorrhage and mild bleeding grouped into bleeding/hemorrhage; pain and tenderness
`grouped into pain] to eliminate clinically unimportant distinctions. Miscellaneous
`cutaneous adverse events [e.g., swollen cheek, warm sensation, herpes simplex, rash (2
`events)], are classified as Skin disorder, NOS.
`
`Both intensity and duration ofan adverse event are relevant in assessing its impact on the
`patient. The duration ofpain (experienced by patient 18106) was 22 minutes; the
`duration oftenderness (in patient 18303) was 6 days. Patients 18106, 18116, 18117, and
`18501 developed ulcers that lasted 2 days, 3 days, 1 day, and 27 days, respectively.
`
`In response to medical reviewer’s request, sponsor submitted clinical slides ofpatients
`who developed ulcers following PDT. As judged fiom the submitted slides, all ulcers
`healed without leaving evidence of scar formation.
`
`Laboratory Evaluations
`
`For patients enrolled in the active treatment arm, mean hematocrit decreased from 44.04
`at screening to 43.67 at Day 2. Compared to the percentage ofpatients in the active
`treatment arm with hematocrit levels below the normal range at screening (6%), more
`patients had abnormally low hematocrit levels at Day 2 (12%). This was higher than the
`percentage ofpatients in the vehicle arm with abnormally low hematocrit levels at Day 2
`(7%). Five patients [18107, 18202, 18223, 18510,.and 18703] had hematocrit levels
`within the normal range at screening that subsequently fell below the normal range by
`Day 2. The mean decrease in hematocrit levels for these five patients was 2. None ofthe
`changes in hematocrit were dramatic or of direct concern clinically: Patient 18704 had
`the lowest hematocrit level among active-treatment patients at Day 2 (value of 32), but
`this value was not substantially lower than what had been observed at baseline (32.3).
`Two of these five patients underwent retreatment,_but only one again experienced a drop
`in hematocrit. The five patients had no common concomitant medications, no common
`
`Reviewer '5 Comment: It would be appropriate toperform a Phase 4 safety study to
`confirm that any treatment-induced decrease in hematocrit is small and not clinically
`relevant.
`
`Two patients in the LEVULAN®-treated arm (nos. 18103, 18106), but no vehicle-treated
`patients, had slightly elevated urinary levels of aminolevulinic acid 'on Day 2. These two
`patients each had six target lesions treated.
`
`Reviewer’s Comments/Conclusions of stud results
`8.3.1.8
`This clinical study convincingly demonstrated that LEVULAN® is effective for
`treatment of multiple actinic keratoses ofthe face. Among patients with scalp lesions, a
`
`

`

`38
`
`higher percentage ofpatients in the active treatment completely cleared all lesions
`compared to the patients in the vehicle treatment, but the difference did not reach
`statistical significance. The likely reason for this failure to reach statistical significance is
`that comparatively few subjects with scalp lesions were enrolled in this trial. Within the
`limited time frame of this trial, the vast majority of lesions that achieved complete
`remission remained in remission. Re-treatment with LEVULAN®/blue light at week 8 of
`those lesions not in complete remission resulted in a significant increase in the complete
`clearance rate at week 12. Treatment is more effective for thinner lesions.
`
`The preponderance of recorded adverse events were cutaneous, as would be expected for
`a topical treatment in which limited systemic absorption of the drug occurs. During and
`shortly after treatment, erythema and buming/stinging occurs for every patient, usually in
`most of the target lesions. Half the LEVULAN®/blue light treated patients experience
`edema at the target lesions. These adverse events largely resolved within a week afler
`treatment. Other commonly experienced (>5%) cutaneous adverse events included
`scaling/crusting, itching, ulceration, hypo/hyperpigmentation, erosion, and miscellaneous
`cutaneous disorders. These events were predominantly mild to moderate in severity and
`short-lived. The only noteworthy treatment-emergent laboratory abnormality was a
`decrease in hematocrit, observed in a subset of the patients, that was not clinically
`significant.
`
`8.3.2 Trial #2—ALA-079 ~
`8.3.2.1
`Objective/RationaIe/Design
`Identical to ALA-018.
`
`8.3.2.2
`Protocol Overview
`All aspects of this protocol, including procedures, evaluability criteria, defined endpoints,
`and statistical considerations, were identical to ALA-018. ALA-019 was conducted at 8
`centers. 126 patients were randomized: 93 to receive LBVULAN® and blue light, and
`33 to receive vehicle and blue light. Four patients in the LEVULAN® aim and 2 patients
`in the vehicle arm discontinued from the study.
`
`AWEARS ““3 WAY
`0N ORlGlNAL
`
`.,
`
`-_,
`
`_
`
`

`

`39
`
`
`
`
`Patient Discontinuations: ALA-019
`Patient Number
`Treatment Ann
`
`19102
`
`LEVULAN®
`
`
`
`Reason for
`
`Patient Non-
`
`
`
`Time of Last
`
`Week 8
`
`19107
`
`LEVULAN®
`
`Week 12
`
`
`
`
`
`
`Patient
`
`
`Discontinued (out of
`
`town)
`
`
`Patient Death (due Week 4
`to widespread
`
`
`
`
`carcinoma)
`_-_—
`Patient Non-
`
`
`--__
`
`
`--——
`
`
`
`
`
`
`
`19702
`
`19108
`
`19714
`
`LEVULAN®
`
`LEVULAN®
`
`Vehicle
`
`Vehicle
`
`
`
`.
`
`..
`
`Patient
`
`Patient
`
`-0
`
`
`
`
`
`
`
`Week 9
`
`Week 9
`
`Week 1
`
`APPEARS nus WAY
`on ORIGINAL
`
`

`

`
`
`
`
`
`
`
`
`
`
`
`
`40
`
`
`
`
`
`
`
`
`
`
`Study Results
`8.3.2.3
`8.3.2.3.1
`Demographics, Evaluabilig
`
`Summa
`of Baseline Demo ra-
`
`_—_m__p-value‘
`amm—
`um_————
`_—2__—-a-
`”mas—m—
`m—mm—
`_——_—
`mum-Inna.
`-__—-—
`Em—————
`__————
`_———_
`___—_
`_—_—-
`_——_—
`m__—_—
`____—
`_———mm-
`_————
`_————
`————-
`Izm——_——
`__———
`m___——
`—_—_—
`_—n:——--m——
`—————
`—————
`_————
`____—
`-——_——
`__—__
`__—_—
`———__
`_——_mm-
`__—_——
`__——-
`Note: Percentages were calculated based on the number of patients with non-missing values in each treatment group.
`' P-value is based on ANOVA with treatment for age and Cochrm-ManteI-Haenszel general association test for sex and
`skin type.
`‘
`' Skin Type:
`l: White; always bums easily; shows no immediate pigment darkening reaction (IPD); never tans.
`H: White; always bums easily; trace [PD; tans minimally and with difficulty.
`"I: White; bums minimally; 1PD+; tans gradually and unifomrly (light brown).
`IV: Light brown; bums minimally; [Pm-4|»; always tans well (moderate brown).
`V; Brown; rarely bums; lPD-H-r; tans profusely (dark brown).
`Vl:
`Dark brown or black; never bums; IPD+-H; tans profusely (black).
`Data Source: End-of Text-Tables 2, 3, and 4; Patient Data Listing 1. 3. and 4; and CRF pages 3. 5, and 7.
`
`Source: Tables ”2.1. “.22. Vol L60. pp. 7-9587, 7-9589
`
`
`
`
`
`
`
`
`
`Reviewer 's Comment: There are no significant difl'erences between the patients
`randomized to LEVULAN® or vehicle treatment. The demographic and baseline
`characteristics ofthe patients in ALA-019 and ALA-018 are quite similar in all respects.
`Some slight differences are that a higher percentage ofpatients in ALA-019 have skin
`type 1, they have more lesions perface and scalp, and a higher proportion ofenrolled
`patients with scalp versusface lesions.
`
`

`

`Second
`First Treatment
`-- Second
`--—-
`I'IT
`Evaluated
`Received/
`Received/WW-
`I'I'I“ Evaluated
`
`
`
`Evaluability b Center/Investi-ator: ALA-O19
`LEVULAN®
`First Treatment
`
`Investigator
`
`D. CHEN
`Chicago, IL,
`USA.
`
`J. FOWLER
`Louisville, KY,
`U.S.A
`
`7
`
`l/2-t-
`
`12
`
`.
`
`1/1
`
`41
`
`2/2
`
`4/4
`
`5/5
`
`2/3
`
`3/4
`
`4/4
`
`3/3
`
`
`
`
`
`
`
`
`
`L. HRUZA
`St. Louis, MO,
`USA.
`
`15
`
`15
`
`6/6
`
`5
`
`5
`
`5
`
`_-H--—.-
`
`12
`
`12
`
`12
`
`3/4
`
`4/4
`
`T. PHILIPS
`
`.
`
`T. RALLIS
`Salt Lake City, UT,
`USA
`
`D. TASHHAN
`
`C. TAYLOR
`
`12
`
`12
`
`10/10
`
`1/2
`
`WA ““_-—I-
`—n---—I-
`"
`‘
`" nn- -I-
`
`
`
`‘l'I'l': patients who are enrolled, randomized, and who receive LEVULAN® or vehicle
`
`4-Patient 19103, who missed visit 8, did not have 100% CR at weeks 4 or 12, and therefore was classified eligible
`Source: Vol. 1.60. Table 6.1 (pg. 7-9540) and Table 10.1.1, (pg. 7-9580) Data Listing 7
`
`
`
`
`
`
`Reviewer '3 Comment: The PDTResponse presumably was not unbearablypainful or
`unpleasant, as 91% (32/35) ofthe patients receiving LEVULAN®/blue light who were
`eligiblefor retreatment at week 8 were Willing to undergo a second round oftreatment.
`
`Diana Chen, M.D. is'idenfified asme the blinded and unblinded investigator for Center
`1. This is a protocol violation. The other seven unblinded investigators were not
`physicians.
`o
`
`Reviewer '3 Comment: The quality ofthe safety data is rendered suspect because most of
`the unblinded investigators are not physicians. At the time ofcompletion ofthis review,
`information about the qualifications ofthe un'blinded investigator was not available. For
`Phase 4 studies to collect additional safety data, Agency should specijy that both
`unblinded and blinded investigators are physicians.
`
`APPEARS THIS WAY
`0" ORIGINAL
`
`

`

`8.3.2.3.2 Primary Efficacy Results
`
`42
`
`
`
`
`Total
`
`59/93
`
`4/32
`
`
`
`
`
`
`I
`
`'
`
`<.001
`
`59/93* 4/330
`
`
`
`
`<.001
`
`.— ((3%) -- ((3%)
`(2%.) - ((0%)
`(0%) - (46%)
`
`Response Rates at Week 8, ITT, L.O.C.F.: ALA-019
`Sponsor Analysis: 100% Complete
`Agency Analysis: 100% Complete
`Response Rate (Tables 22.5, 22.6, 22.7)
`Response Rate
`Active Vehicle
`95%
`p-
`Active Vehicle
`95%
`
` 'U
`
`'
`Confidence
`value
`Confidence
`value
`
`
`
`Interval of
`Interval of
`
`
`
`difference
`difference
`(2%) -
`36%~66%
`36%-66%
`(21%) -
`(0%)
`, -
`
`4/20A
`
`28%-71%
`
`<.001
`
`0/13
`
`Face
`
`47/67
`
`4/19
`
`28%-71%
`
`<.001
`
`47/67
`
`12/26
`
`0/13
`
`-27%-65% <.001
`
`12/26
`
`
`
`
`
`W(%%)
`(27%-65%) <.001
`
`
`
`
`* This ratio differs slightly fi'om that calculated by the statistical reviewer (5 8/93) because
`medical reviewer has not excluded from the count of cleared patients #19410, who
`
`
`achieved 100% CR at week 4, and was then lost to follow-up.
`
`
`0 Denominator is 33 in the Agency I'I'I‘ analysis. Patient 19714 is excluded from
`
`sponsor’s LOCF analysis because of withdrawal from study because of patient non-
`
`, -- a
`. :
`s .- week-le—fl-‘heughthereis no efficacy data on this patient, the patient
`
`
`should be included in ITT analysis. Changing the denominator from 32 to 33 will have a
`negligible effect on treatment outcome.
`
`
`
`
`ADenominator was changed fiom 19 to 20, for same reason as listed above.
`
`(1) outcomes ofpatients treatcd’with LEVULAN® and blue light were
`significantly superior’to outcomes of patients treated with vehicle and blue
`light for‘allpatients studied, fer patients 'with facial lesions, and for patients
`
`with‘scalp lesions'
`.__..-_-_
`_ _
`WI
`_
`, .. _
`Patients with scalporjacelesionsreceiying active treatment had outcomes
`statistically superior tothat of vehicletreated patients, but patients with scalp '
`lesions did not respond aswellas did patients with facial lesions. One possible
`explanation for whypatientswith scalp lesions had poorer responses is that the
`mean lesion grade for scalp lesions (1 .57) was significantly greater (p<.01) than
`the medalssmd aged: for £55: lawn-“13.9) Land lesipns with higher grade did
`not respond as-well to treatment (see below).
`treated with LEVULAN® and blue
`light were inferior to those of patients with .facial lesions treated with
`LEVULAN® and blue light [this difference was not statistically significant
`lp=~05 8])
`' "
`'
`
`To reject the possibility that the results from one or a few of the centers drove the
`outcome depicted above, the 100% CR rate across the different study centers was
`examined (as depicted below).
`
`

`

`43
`
`
`
`
`100% CR Rate at Week 8 by
`Center/lnvesti - ator: ALA-01 9
`Laval—m
`
`Ii
`
`were
`
`6/8 (75%)
`
`2/4
`(50%)
`
`11/12 (92%)
`
`0/4
`
`9/15 (60%)
`
`0/5
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`D. CHEN
`Chicago, IL,
`U.S.A.
`
`J. FOWLER
`Louisville, KY,
`U.S.A
`‘
`
`’
`
`. _
`
`L. HRUZA
`St. Louis, MO,
`USA.
`
`T. PHILIPS
`Boston, MA, USA.
`T. RALLIS
`Salt Lake City. UT,
`U.S.A
`
`9/12 (75%)
`
`7/ 12 (58%)
`
`1/4
`(25%)
`0/4
`
`0/4
`
`D. TASHJIAN
`
`pm... CA, us... -
`
`‘
`
`0/10 (0%)
`
`C. TAYLOR
`
`10/12 (83%)
`
`1/4
`105%)
`
`G. WEINSTEIN
`
`7/12 (58%)
`
`0/3_-
`
`
`The only extreme outlying center (Dr. Tashjian’s) has an unusually low 100% complete
`response rate for patients treated with LEVULAN®/blue light, so including the results
`from this center lowers the overall 100% CR. No single center exerts an untowardly
`positive effect on the primary efficacy variable. The consequence of removing the center
`in which the blinded and unblinded investigator was the same individual would have a
`minimal effect on the 100% CR rate (it would drop from 63% to 62%).
`
`The following table shows that among LEVULAN® patients who had achieved 100%
`complete response by week 8, the majority remafiied clear of all their target lesions at
`week 12.
`
`APPEARS THIS WAY
`
`0N ORIGINAL
`
`

`

`Remission Duration in Patients with 100% CR Observed at
`Week 8: ALA-019
`
`Among patients with 100% CR rate
`observed at Week 8*
`
`
`
`LEVULAN(%)(N=58'I*)
`
`
`
`100% CR MAINTAINED AT WEEK 12
`100% CR LOST AT WEEK 12
`
`
`
`
`48 (83%)
`10 (17%)[6 of the
`patients (19101, 19413,
`19506, 19511, 19705,
`19713) had facial lesions
`recurring, 4 (19317,
`
`19320, 19505, 19514)
`had scalp lesions
`recurring]
`
`*The 100% CR rate observed at Week 8 was either achieved at Week
`
`
`
`
`
`
`
`
`4 and maintained, or achieved at Week 8.
`
`*I‘The number of cleared patients counted in this table (58) is less than
`
`that counted in preceding table (59), because patient #19410 (who
`
`
`achieved 100% CR at week 4, and was then lost to follow-up) was
`
`excluded.
`
`
`
`Those subjects with any persistent target lesions at Week 8 were eligible for retreatment
`of those lesions at that time, following the same randomization scheme as in the original
`application. As depicted in the following table, evaluating the population of all patients,
`and the subset of patients with. facial lesions, repeat treatment of persistent target lesions
`with LEVULAN®lblue light converted significantly more patients to 100% complete
`responses by week 12 than did repeat treatment with vehicle/blue light. For the subset of
`patients with persistent scalp lesions, retreatment with LEVULAN®/blue light trended
`toward a benefit, but the difference was not statistically significant (likely due to the
`small numbers of retreated patients).
`5
`
`Among Patients with Lesions Retreated at Week 8, Patients with 100% CR
`Rate at Week 12: ALA-0.19.
`.__1.
`
`
` Vehicle 95%
`
`'
`Confidence
`Interval of ’
`
`
`difference
`
`
`
`
`
`
`
`
`Sources: Table 24.4, 24.5, and 24.6, Vol 1.60, pp. 7-9812 to 14
`
`.026
`
`-021
`.
`
`

`

`~The following table compares the lesion clearance rate for different lesion grades
`(outcomes for lesions of the same grade located on face and scalp are pooled for this
`analysis).
`
`Lesion Complete Response Rate at Week 8, L.0.C.F. for different
`Lesion Grades: ALA-019
`
`LEVULAN® .Yehicle
`'
`
`
`
`81/187
`
`
`
`95% Confidence
`Interval of
`
`Difference
`
`34%—50%
`
`
`
`232/359
`
`.24/111
`
`48%-66%
`
`p-value
`
`<.001
`
`
`
`
`
`366/429 _
`Lesion Grade 1
`(85%)
`(lesions are slightly
`
`
`palpable, and better felt
`
`than seen)
`
`
`
`Lesion Grade 2
`
`(lesions are moderately
`
`
`(79%)
`(22%)
`thick actinic keratoses,
`
`
`
`easily seen and felt)
`
`
`
`Source: Tables 18.5, 18.6
`
`
`While the active treatment outcomes are significantly superior to those of vehicle
`—*****mm each le‘s'ion ~gfade‘, the complete response rate decreases as lesion grade
`increases.
`
`8.3.2.3.3 Secondam Efficacy Results
`
`As with study ALA-018, the protocol-specified secondary efficacy variables included
`investigator’s and patient’s cosmetic evaluation. Grading of cosmetic response of treated
`lesions and of overall response was in four categories: excellent, good, fair, and poor.
`
`Cosmetic Evaluation of Lesion by _
`Blinded Investigator, Week 12: ALA-
`018 '
`
`
`
`Excellent
`Good
`
`
`Fair
`
`
`
`
`Cosmetic Scale
`
`mamm
`
`
`
`70 (25%)
`
`611 (80%)
`79 (10%)
`43 (6%)
`
`27 (4%)
`760
`
`Total
`
`-
`
`P value<.001 (Cochran-Mantel-Haenszel mean
`score test (RIDIT scores)
`‘
`Source: Table 11.4.1.6.1.
`
`
`
`
`
`
`
`
`
`
`

`

`46
`
`Reviewer ’s Comment: The way these categories are definedpermits investigator to
`capture in one grading scale (a) whether treatment has resolved a lesion, and (b) whether
`lesion resolution is accompanied by a cosmetically acceptable appearance. It is
`impossible to teasefrom the cosmetic evaluations the relative weights ofthese two eflects
`in determining the "cosmetic " grade. That better "cosmetic ” responses are observed
`following LEVULAN® treatment than vehicle treatment may reflect the higher likelihood
`oflesion curefollowing LEVULAN® treatment, and may not necessarily reflect
`aesthetically satisfactory outcomes with 'IsEVULAN® treatment.
`
`8.3.2.3.4 Safety
`Extent of Exposure
`93 subjects received one treatment of LEVULAN®lblue light and 31 subjects received
`two treatments of LEVULAN®/blue light
`
`Discontinuations
`,,
`No subjects were permanently or temporarily discontinued from the study due to
`laboratory abnormalities. No patients discontinued due to adverse events experienced
`during the light treatment.
`
`Adversefivents—""“" '
`
`Local cutaneous adverse events (i.e., photodynamic response) were reported separately.
`Sponsor reports that 38 patients on active treatment and 12 patients receiving vehicle
`treatment experienced adverse events during this study. Sponsor’s table below lists the
`number of patients with clinical adverse events, by body system and COSTART
`terminology. If a patient experienced more than 1 episode of an adverse event, the patient
`was counted only once for that event. If a patient had more than 1 adverse event in a
`body system category, the patient was counted only once in that body system total.
`
`/
`
`APPEARS THIS WAY
`0N ORIGINAL
`
`

`

`47
`
`
`
`No. (%) of Patients with Clinical Adverse Events by Body
`System: ALA-O1 9
`Body System Category/
`Adverse Event (COSTART)
`
`LEVULAN®
`
`Velncle
`
`
`
`(N=33)
`CN=93)
`—_
`
`wm: Anymverseavems
`
`Boayasawrmle
`Momma
`
`Accidenmmjury
`cuestpatn
`raceEdema
`Headache
`
`Infection
`Neckram
`
`cardiovascumsystem
`Bradycardia
`Hypertension
`
`Digestive System
`'
`'
`Carcinoma of Liver
`
`_—
`
`——
`
`(
`0 0%)
`l (1%)
`0 (0%)
`3 (3%)
`__
`
`(
`
`Gastrointestinal Disorder
`Jaundice
`
`0 (0%)
`l (1%)
`l (1%)
`
`Hernic and Lymphatic System
`Blood Dyscrasia
`
`Periodontal Abscess
`l (1%)
`Rectal Disorder —
`0 (0%)
`Tooth Disorder
`1 (1%)
`_—
`(
`I (1%)
`l 3%)
`0(0%)
`0 (0%)
`l (1%)
`——
`1 (1%)
`0 (0%)
`l (1%)
`O (0%)
`——
`(
`
`l (3%)
`0 (0%)
`0 (0%)
`0(0%)
`1 (3%)
`
`
`
`
`
`
`
`’
`
`(
`
`(continued)
`
`APPEARS THIS WIN
`0" ORIGINAL
`
`2;g0
`
`Metabolic and Nutritional System
`Creatinine Increased
`
`'
`
`'E
`
`Musculoskeletal System
`Anhralgia
`
`Myal'gia
`Tendon Disorder
`
`
`
`

`

`48
`
`Table
`No. (%) of Patients with Clinical Adverse Events by Body
`8 stem
`
`Body System Ca‘tegory/
`
`LEVULAN®
`
`Vehicle
`
`.m———
`__-M-
`
`—_—
`
`”_—
`
`
`
`
`
`
`
`
`«_—
`_—_
`
`-_——I§m—
`
`
`
`
`
`Emma—_—
`_—_
`mama—_—
`———
`
`_—
`
`Data Source: End-of Text Table 31; Patient Data Listing 13; and CRF page 41.
`
`
`
`
`The six patients in the LEVULAN® arm who experienced serious adverse events
`(patients 19101, 19214, 19410, 19511, 19613, and 19802) experienced basal cell
`carcinoma (site unidentified), squamOus cell carcinoma (right lower leg), liver failure and
`death (see section 10.1.1), pneumonia, automobile accident, and bradycardia.
`
`

`

`BEST POSSIBLE copy
`
`49
`
`Investigators considered these adverse events unrelated or remotely related to treatment,
`as does the medical reviewer.
`
`Adverse Events: pre-PDT
`
`42% of active treatment patients develop burning/stinging between Baseline A and B,
`while 6% of vehicle treatment patients develop burning/stinging; 10% of active treatment
`patients develop edema between Baseline A and B, while no vehicle treatment patients
`develop edema between Baseline A ath.
`
`Reviewer '3 Comment: These results are either attributable to irritancy or to an
`inappropriate photodynamic response. Since no irritancy study has been performed with
`the to-be-marketedfonnulation, it is not possible to discern which ofthese two
`alternative explanations is correct.
`
`Adverse Events: peri-PDT
`
`The medical reviewer analyzed the prevalence of signs and’or symptoms expected during
`a photodyna