Cancer Chemother Pharmacol (1990) 25: 395-404
`
`~
`
`ancer
`hemotherapy and
`harmacology
`© Springer-Verlag 1990
`
`A strategy for the development of two clinically
`active cisplatin analogs: CBDCA and CHIP
`
`Brenda J. Foster1, Bonnie J. Harding1, Mary K. Wolpert-DeFilippes2, Lawrence Y. Rubinstein3, Kathleen Clagett-Carr',
`and Brian Leyland-Jones1
`
`1 Investigational Drug Branch, 2 Developmental Therapeutics Program, and 3 Biometric Research Branch, National Cancer Institute,
`Bethesda, Maryland, USA
`
`Summary. The antitumor agent cisplatin has a broad an(cid:173)
`titumor spectrum and has been incorporated into regimens
`that are curative for some malignant diseases. However,
`one of the major limitations to its clinical usefulness is the
`incidence of severe toxicities involving several major
`organ systems. Therefore, much enthusiasm has been
`generated for the development of cisplatin analogs that
`demonstrate an improved therapeutic index in some pre(cid:173)
`clinical models. The two most promising analogs are
`CBDCA (carboplatin) and CHIP (iproplatin). The preclin(cid:173)
`ical and early clinical trial results have demonstrated that
`these two compounds show activity in cisplatin-responsive
`tumors. The preclinical background providing the ration(cid:173)
`ale for the clinical development of these two analogs is
`described. We suggest a means of screening for each
`analog's clinical antitumor activity and determining the
`analogs' utility against specific malignant diseases com(cid:173)
`pared with that of the parent compound or standard treat(cid:173)
`ment.
`
`Introduction
`A report hy Rosenherg et al. [52] describing the antitumor
`activity of platinum compounds led to wide-scale clinical
`investigations of these and other platinum coordination
`complexes. From these clinical studies, a role for cisplatin
`in the treatment of a variety of neoplasms was established
`[34]. The severity of the gastrointestinal and renal toxicities
`associated with cisplatin administration encouraged trials
`with schedule manipulations, antiemetic regimens, hydra(cid:173)
`tion schema with and without diuretics, and renal prophy(cid:173)
`laxis such as hypertonic saline and thiosulfate. In addition,
`interest was stimulated in the development of alternative
`platinum compounds with a better therapeutic index and a
`similar or improved antitumor activity spectrum.
`Preliminary results against Ll210 leukemia and sar(cid:173)
`coma 180 in mice [52] demonstrated that the most effica(cid:173)
`cious platinum compounds had either a cis configuration
`for the chloride groups [platinum(II) coordinated com-
`
`Offprint requests to: Brian Leyland-Jones, National Cancer In(cid:173)
`stitute, Cancer Therapy Evaluation Program, Investigational
`Drug Branch, Executive Plaza North, Room 731, Bethesda, MD
`20892, USA
`
`plexes] or were platinum (IV) coordinated complexes. The
`three properties required for platinum compounds to have
`antitumor activity are: (a) neutrality; (b) possession of a
`pair of cis leaving groups that have a !ability similar to that
`of the chlorides; and (c) possession of ligands other than
`the leaving groups [9, 11, 51]. Two cisplatin analogs with
`these structural characteristics, CBDCA [diammine 1,1
`cyclobutane dicarboxylato Pt(II), JM-8, NSC-241240] and
`CHIP
`[bis-isopropylamino-trans-dihydroxy-cis-dichloro
`Pt(IV), JM-9, NSC-256927], are shown in Fig. I. Both are
`undergoing clinical trials sponsored by the National Can(cid:173)
`cer Institute (NCI). This paper provides a brief review of
`the preclinical and phase I data on CBDCA and CHIP to
`present the background for the development of two first(cid:173)
`generation platinum coordination complexes and then
`describes the NCl's planned development of these two
`agents.
`
`Mechanism of action
`Platinum coordination complexes inhibit tumor growth by
`their effects on DNA replication. The binding of these
`complexes to DNA is similar to that of hifunctional
`alkylating agents and has been shown to correlate with
`cytotoxicity in intact cells [15, 41, 42, 64]. All platinum(II)
`analogs (including CBDCA) induce DNA shortening and
`superhelical conformational changes, whereas plati(cid:173)
`num(IV) compounds (including CHIP) produce DNA de(cid:173)
`gradation [40].
`Guanine residues have been shown to he a site of DNA
`cross-linking [26, 32, 36, 54]. The kinetics of the cisplatin(cid:173)
`DNA cross-link formation in Ll210 leukemia, previously
`reported by Zwelling et al., required 12 h drug incubation
`for maximal cross-link formation. For the much less
`cytotoxic trans isomer, maximal cross-linking occurred hy
`the end of 1 h drug incubation [63]. Other investigators
`have also reported differences in DNA-protein cross-link
`kinetics between the cis and trans isomers [35, 37, 41,
`42, 54].
`Although both CBDCA and CHIP have been shown to
`react with DNA [8, 20, 40], Mong et al. [40] reported dif(cid:173)
`ferences in the types of changes induced in PM-2 DNA by
`these agents. Cisplatin and CBDCA, both platinum(II)
`compounds, produced alterations in tertiary DNA confor(cid:173)
`mations but had little effect on linear PM-2 DNA; indeed,
`superhelical structure was a prerequisite for their cyto(cid:173)
`toxicity. The activity of both compounds was inhibited by
`
`NOVARTIS EXHIBIT 2083
`Breckenridge v. Novartis, IPR 2017-01592
`Page 1 of 10
`
`

`

`396
`
`CHEMICAL STRUCTURE
`
`COMMON NAME
`
`NSC #
`
`JM#
`
`NH3.......___
`
`/Cl
`
`Pt
`
`NH3/
`
`'----c1
`
`NH3"'
`
`Pt
`
`NH3/
`
`0
`II
`/0 - C )< )
`
`"-o-c
`II
`0
`
`CH3
`I
`OH
`CH 3 - CH - NH2.......___ I
`/Cl
`Pt
`'----c1
`I
`OH
`
`CH - CH - NH /
`3
`2
`I
`CH3
`
`Cis Platinum,
`Cisplatin
`
`119875
`
`CBDCA,
`Carboplatin
`
`241240
`
`8
`
`CHIP,
`Iproplatin
`
`256927
`
`9
`
`Fig. 1. Structures, names, and NSC numbers of cisplatin and two analogs
`
`Table J. Antitumor activity of cisplatin, CBDCA, and CHIP against. the tumor panel
`
`Tumor system
`
`CBDCA:
`
`Treatment Cisplatin:
`schedule
`(i.p.)
`
`Dose
`rangeb
`(mg/kg)
`
`T/C±SE• Score• Dose
`(%)
`range
`(mg/kg)
`
`T/C±SE Score
`(%)
`
`Murine tumors:
`qld,
`i.p. B16
`melanocarcinoma days 1-9
`q7d,
`s.c.CD 8 F,
`mammary tumor
`days 1-29
`q7d,
`s.c. colon
`days 2,9
`38 tumor
`q7d,
`i.p. Ll210
`days 1-9
`leukemia
`qld,
`i.v. Lewis
`days 1-9
`lung-carcinoma
`
`Human tumor xenografts:
`q4dx3,
`s.c. CX-1
`days 14-22
`colon tumor
`q4dx3,
`s.c. LX-1
`days 14
`lung tumor
`q4dx3,
`s.c.MX-1
`days 14
`mammary tumor
`Optima i.p. dose, days 1-9
`
`2-4
`
`(81 ±8)
`
`12.5- 50.0
`
`(63)
`
`2-8
`
`(69)
`
`4-8
`
`(3)
`
`++
`
`50.0
`
`25.0
`
`(140)
`
`(43)
`
`25.0
`
`25.0
`
`25.0
`
`(41)
`
`(94)
`
`(59)
`
`1.6mg/kg
`
`16 mg/kg
`
`14mg/kg
`
`• Antitumor activity expressed as the mean optimal TIC(% indicated) (NIH Publication 84 2635)
`b Dose range for which optimal activity in a dose response was observed. Minimal criteria for activity: % T/C for survival assays -
`£1210, B/6, ;;,, 125%; Lewis lung,;;,, 140%; %TC for tumor weight-inhibition assays- CD 8 F1, colon 38, ..;42%; CX-1, LX-1, MX-1, ..;20%
`0 DN2 criteria for activity:% T/C for survival assays, ;;,, 150%; % TIC for tumor weight-inhibition assays, ..; 10% (values in parentheses).
`+ +, Minimal criteria for activity; - , no activity
`
`CHIP:
`
`Dose
`range
`(mg/kg)
`
`T/C±SE Score
`(%)
`
`12.5
`
`50.0
`
`25.0
`
`166
`
`(6)
`
`(46)
`
`++
`
`++
`
`12.5-25.0 183±14 ++
`
`0.2-4.0
`
`178±2
`
`++
`
`12.5- 25.0 172± 4 ++
`
`4.0-12.5
`
`(1 ±1) ++
`
`50.0-100.0
`
`(8)
`
`++
`
`2.0-16.0
`
`(38±5)
`
`2.0- 4.0 162±2
`
`0.5- 2.0 153±6
`
`+
`
`++
`
`++
`
`100.0-200.0
`
`(33± 11)
`
`25.0- 64.0
`
`148± 7
`
`+
`
`+
`
`6.3- 25.0 119± 7
`
`6.3-12.5 129
`
`NOVARTIS EXHIBIT 2083
`Breckenridge v. Novartis, IPR 2017-01592
`Page 2 of 10
`
`

`

`Table 2. Comparative activity of cisplatin, CBDCA, and CHIP against mouse leukemias
`
`397
`
`Reference
`
`Activity
`
`CBDCA:
`
`Dose
`(mg/kg)
`
`Activity
`
`CHIP:
`
`Dose
`(mg/kg)
`
`Activity
`
`157%-186% TIC
`164%-229% TIC
`157%-285% T/C
`
`32
`128
`64
`
`171%T/C
`150%T/C
`157%T/C
`
`50
`32
`16/day
`
`137%T/C
`171%T/C
`207%T/C
`
`(2, 7,8,41,
`45,46]
`
`Cisplatin:
`
`Dose
`(mg/kg)
`
`4-10
`8
`2/day
`
`Tumor
`
`Treatment
`schedule
`
`LJ210
`
`Day!
`Day 1
`Days 1-9
`Days 1-9
`or
`Days 1, 5, 9
`Ll210/CDDP Day l
`LJ210
`Day l
`in vivo (cid:157)
`in vitro
`P388
`
`Days 1-9
`Days I, 5, 9
`
`1.6-2.4/ day
`
`186%-257%T/C
`
`25/day
`
`152%T/C
`
`25/day
`
`191% TIC
`
`4-8
`9
`
`94%-l31%T/C
`Surviving
`fraction
`= 50%•
`
`120
`336
`
`25
`
`113%T/C
`Surviving
`fraction
`= 50%•
`152%T/C
`
`32
`135
`
`18
`50
`
`118%T/C
`Surviving
`fraction
`= 50%•
`202%T/C
`154%T/C
`
`[46]
`[27]
`
`[7, 8]
`
`• In vitro colony formation assay. Shown is the dose that caused a 50% reduction in the colony formation of tumor cells in vitro following
`treatment of tumor-bearing mice. o/o TIC, Median survival time of drug-treated tumor-bearing mice compared with that of mice treated
`with vehicle only. Drugs were given i.p.
`
`sodium chloride. CHIP, a platinum(IV) compound,
`caused breakage of covalently closed, circular PM-2
`DNA; this breakage was not inhibited by sodium chloride.
`This suggests involvement of the axial trans bonds rather
`than the equatorial cis bonds (40]. In addition, the con(cid:173)
`centration of CHIP required to produce DNA damage was
`higher than that required for cytotoxicity [40], suggesting
`that DNA breakage may not be the primary mechanism of
`cytotoxicity.
`
`Antitumor activity
`CBDCA and CHIP have been tested for antitumor activity
`against many in vitro and in vivo tumor models, including
`human tumor xenografts. Comparative results obtained
`with the analogs and cisplatin at optimal doses against
`tumors used in a preclinical screen at the NCI are shown
`in Table I [60, 61]. These data are the results of screening
`carried out under the auspices of the Developmental
`Therapeutics Program (Division of Cancer Treatment,
`NCI, Bethesda, Md). Cisplatin showed the broadest ac-
`
`Table 3. Toxicity of cisplatin, CBDCA, and CHIP after a single
`i.v. dose in male F344 rats
`
`Cisplatin
`CHIP
`CBDCA
`mg/kg (mg/m 2) mg/kg (mg/m 2) mg/kg(mg/m 2)
`
`6 (36)
`8 (48)
`
`1.3
`
`LD10
`LDso
`LDso"
`LD 10
`LD5ob
`LD50
`
`52.5 (313.2)
`60.9 (365.4)
`
`33.4 (200.4)
`39.0 (234.0)
`
`1.2
`
`7.6
`
`1.2
`
`4.9
`
`LD 10 or LD 50 is the dose that produced lethality in I 0% or 50%,
`respectively, of the rats treated ( data from [581)
`• LD 50 compound in mg/kg
`. .
`.
`/k = tox1C1ty quotient
`d .
`LD
`10 compoun tn mg g
`.
`b LD 50 analog in mg/kg
`LD 50 cisplatin in mg/kg = potency ratio
`
`Table 4. ( a) Comparative toxicity of cisplatin, CBDCA, and CHIP
`after a single i.v. injection in male F344 rats
`
`Parameter
`
`Cisplatin
`
`CBDCA
`
`CHIP
`
`Hematocrit
`WBC
`BUN
`Creatinine
`SGPT
`Body weight loss
`Histopathology:
`Renal
`Lymphatic
`Hematopoietic
`Gastrointestinal
`Total score:
`
`I
`3
`3
`3
`l
`3
`
`4
`4
`3
`4
`30
`
`3
`2
`l
`I
`1
`1
`
`I
`1
`4
`l
`16
`
`2
`3
`1
`I
`1
`2
`
`3
`4
`3
`1
`21
`
`(b) Scoring used for comparative toxicity of platinum compounds
`after single-dose administration
`
`Parameter
`
`Scoring system and definitions
`
`Hematocrit,
`WBC
`
`I = < 20% decrease
`2 = 20%-50% decrease
`3 = > 500/4 decrease
`BUN, creatinine, I = <50% decrease
`2 = 50%-200% increase
`SGPT
`3 = > 200% increase
`Body weight loss I = no weight loss (maybe slowing of growth)
`2 = < l0% (or < 15% serial bleeding) weight loss
`3 = ~ I 0% ( or > 15% serial bleeding) weight loss
`1 = no lesions
`2 = mild lesions in few animals
`3 = lesions of moderate.to marked severity
`4 = lesions of marked to extreme severity
`
`Histopathology
`
`WBC, leukocyte count; BUN, blood urea nitrogen; SGPT, glu(cid:173)
`tamic pyruvic transaminase
`Data from [58]
`
`NOVARTIS EXHIBIT 2083
`Breckenridge v. Novartis, IPR 2017-01592
`Page 3 of 10
`
`

`

`398
`
`tivity spectrum, with significant activity against i. v. Lewis
`lung carcinoma and s. c. human mammary xenograft [60,
`61], neither of which were affected by CBDCA or CHIP.
`Both cisplatin and CBDCA showed a similar level of ac(cid:173)
`tivity against s. c. colon 38, whereas CHIP showed no ac(cid:173)
`tivity. Cisplatin and CHIP showed quantitatively better
`activity against i. p. Ll210 than did CBDCA [60, 61).
`The results of comparative experiments in mouse
`leukemias are summarized in Table 2 [4, 9, 10, 29, 45, 49,
`50, 58). The L1210 in vivo and in vitro results clearly indi(cid:173)
`cate that cisplatin has the highest potency, followed by
`CHIP, with CBDCA being the least potent [29]. An L1210
`line made resistant in vitro to cisplatin (Ll210/CDDP)
`demonstrated cross-resistance to CBDCA and CHIP (49].
`
`Toxicology
`Comparative toxicologic studies showed CBDCA and
`CHIP to be less potent than the parent compound, as
`evidenced by the defined toxic doses shown in Table 3 [58).
`The severity of myelosuppression, nephrotoxicity, and
`gastrointestinal toxicity caused by the parent compound
`was qualitatively different from that observed after treat(cid:173)
`ment with the two analogs, as shown in Table 4 [29, 45, 50,
`58). Both CBDCA and CHIP produced more hematologic
`toxicity than did cisplatin, but they caused much less renal
`toxicity than the parent drug. Cisplatin produced more
`severe histopathologic lesions in the gastrointestinal tract
`than did either analog.
`In summary, toxicologic studies showed the two
`analogs to be less potent than cisplatin, and, although the
`same organ systems (hematologic, renal, and gastrointes(cid:173)
`tinal) were affected by all three compounds, the patterns
`of toxicity were different. The analogs consistently showed
`less renal and gastrointestinal toxicity but more hema(cid:173)
`topoietic toxicity than did cisplatin.
`
`Clinical studies results
`
`Phase I trials
`Comparative results from phase I studies of cisplatin,
`CBDCA, and CHIP in adults are shown in Table 5 [5-7,
`12-14, 17, 22, 24, 25, 27, 31, 33, 46, 47, 53, 55, 57, 59).
`Based on the total dose (in milligrams) tolerated for each
`drug, cisplatin is the most potent; CHIP, intermediate;
`and CBDCA, the least potent. CBDCA and CHIP differed
`from cisplatin in the relative severity of their gastrointes(cid:173)
`tinal, neurologic, renal, and hematologic side effects.
`Hematologic effects, especially thrombocytopenia, were
`dose-limiting for CBDCA and CHIP (5, 6, 12, 13, 17, 22,
`24, 27, 31, 46, 47, 53, 55, 57, 59], whereas renal,
`hematologic, and gastrointestinal effects were frequently
`dose-limiting for cisplatin [12, 22, 53, 57]. Diarrhea was
`reported from studies of CHIP, but it was not dose-limit(cid:173)
`ing (5, 13, 17, 24, 47]. Renal toxic effects observed in
`studies of CBDCA and CHIP occurred in patients who
`had preexisting renal disease or a concomitant nephro(cid:173)
`toxic event [6, 14, 17, 27, 47]. No new neurologic toxicity
`was found with administration of the analogs; however,
`exacerbations of preexisting neurologic defects were ob(cid:173)
`served following treatment with CBDCA [6, 14, 27, 55].
`Antitumor effects were reported from the phase I trials of
`each compound, particularly in patients with ovarian car(cid:173)
`cinoma. In summary, less renal toxicity was seen with the
`
`analogs and hematologic toxic effects were dose-limiting
`in phase I testing of CBDCA and CHIP, confirming the
`results seen in preclinical toxicologic studies.
`
`Clinical pharmacokinetics
`
`The clinical pharmacokinetic parameters of the three com(cid:173)
`pounds after i. v. single-dose administration are sum(cid:173)
`marized in Table 6. Total and filterable (free, non-protein(cid:173)
`bound) platinum values were determined using flameless
`atomic absorption spectrophotometry [18, 19, 21, 24, 43].
`Following CBDCA or CHIP administration, the plot of
`the plasma levels for either total or filterable platinum was
`most often described as biexponential. The initial half-life
`(t112) was usually < I h, whereas the terminal half-life (t112 P)
`ranged from 7 h to over 5 days. This biexponential pattern
`was not reported for cisplatin. Thus far, no major phar(cid:173)
`macokinetic differences have been observed that explain
`the differences in clinical potency and toxicity of these
`three analogs.
`
`Developmental plans
`The simultaneous clinical development of CBDCA and
`CHIP has stimulated many questions regarding the rela(cid:173)
`tive utility of each with respect to the other as well as to
`cisplatin. The scientific questions center around the rela(cid:173)
`tive therapeutic index (antitumor effects vs acute and
`chronic toxic side effects) of each compound relative to
`the others. This section describes some of the clinical
`developmental plans for these two analogs as well as
`giving specific illustrative examples for each of the three
`main disease categories.
`
`Disease-oriented strategy. To incorporate the concept of
`relative therapeutic index into the phase II and phase III
`developmental plans, diseases were divided into three
`major categories according to cisplatin responsiveness and
`whether or not cisplatin was an important component of
`currently used standard treatment of the advanced disease.
`Illustrative examples of these disease categories are given
`in Table 7 and include the following:
`
`A Cisplatin-sensitive diseases, where standard therapy in(cid:173)
`corporating cisplatin is curative; examples include germ(cid:173)
`cell tumors and epithelial ovarian carcinomas. In this
`category, it is highly likely that CBDCA and CHIP would
`have some antitumor activity; in fact, hints of tumor
`responsiveness were seen in patients with ovarian car(cid:173)
`cinoma entered in the phase I trials. In this category, the
`usefulness of a traditional phase II trial was questioned. A
`phase II trial entering 30-40 patients would delineate an
`analog's antitumor activity with such broad confidence
`limits that it would not be possible to determine the ac(cid:173)
`tivity relative to that of the parent compound. Therefore,
`the plan was to move directly from phase I testing to phase
`III comparative trials.
`An illustrative example for this category is provided by
`a comparative trial of one analog with the parent com -
`pound. Patients with advanced ovarian carcinoma who
`had not received prior chemotherapy were randomized to
`receive a combination of either CBDCA plus cyclophos(cid:173)
`phamide or cisplatin plus cyclophosphamide [l]. The
`cyclophosphamide dose (mglm2) was the same in each
`combination. Preliminary results show equivalent activity;
`
`NOVARTIS EXHIBIT 2083
`Breckenridge v. Novartis, IPR 2017-01592
`Page 4 of 10
`
`

`

`Table S. Comparative adult phase I studies of cisplatin, CBDCA, and CHIP
`
`Schedule
`
`Cisplatin:
`
`CBDCA:
`
`CHIP:
`
`Maximal Major toxicities
`dose(s)
`each day Dose-
`limiting
`(mg/m 2)
`
`Others
`
`Single dose
`
`200, 100
`
`Renal
`Nausea&
`vomiting
`
`RBC, WBC, pits
`Hearing, loss, tinnitus
`Hyperuricema
`
`Twice weekly
`x2-4week
`
`Dailyx 5
`
`15, 60
`
`WBC, RBC, pits Renal
`Nausea&
`Tinnitus
`vomiting
`40,24, 15 Renal
`
`Nausea&vomiting
`iRBC, WBC, pits
`Tinnitus, hearing loss
`Heart failure with
`conduction defects
`
`Maximal Major toxicities
`dose(s)
`each day Dose-
`limiting
`(mg/m 2)
`
`Others
`
`!pits
`
`520,
`550,
`440,
`600
`
`Nausea&
`vomiting
`!WBC,RBC
`Renal
`Malaise
`Neuropathy
`
`Refer-
`ence
`
`[4, 24,
`26, 29)
`
`Maximal Major toxicities
`dose(s)
`each day Dose-
`limiting
`(mg/m 2)
`
`Others
`
`350,
`350
`
`iplts
`
`iWBC, RBC
`Nausea & vomiting
`Diarrhea
`Hypersensitivity
`(rash)
`
`Refer-
`ence
`
`[3, II]
`
`Refer-
`ence
`
`[21,50)
`
`[5, 53]
`
`[21, 50,
`53]
`
`125,99
`
`iWBC, Nausea&
`pits
`vomiting
`iRBC
`Renal
`Paresthesias
`Myalgia,
`arthralgia
`Renal
`
`iplts
`
`[54]
`
`65,45
`
`iplts
`
`[42]
`
`95
`
`iplts
`
`[16,43]
`
`iWBC,RBC
`Renal
`Nausea&vomiting
`Diarrhea
`Hypersensitivity
`(rash)
`
`iWBC,RBC
`Nausea & vomiting
`Diarrhea
`
`[23]
`
`Weeklyx4
`
`55
`
`iWBC,plts
`
`-80
`
`Bolusq4d
`until toxicity
`24-h continuous -
`infusion
`
`iWBC, pits
`renal
`
`Renal
`Nausea& vomiting
`Tinnitus
`Hypersensitivity
`Nausea & vomiting
`Hearing loss
`
`[10]
`
`150
`
`[31]
`
`RBC, red blood cells; WBC, white blood cells; pits, platelets; t decreased; Mg, serum magnesium
`
`500,
`320
`
`iplts
`
`[12,29]
`
`Nausea&
`vomiting
`Hearing loss
`iMg
`!RBC
`Renal
`
`NOVARTIS EXHIBIT 2083
`Breckenridge v. Novartis, IPR 2017-01592
`Page 5 of 10
`
`

`

`400
`
`Table 6. Clinical pharmacokinetic characteristics determined by atomic absorption spectrophotometry
`
`Cisplatin:
`
`CBDCA:
`
`Total
`
`Filterable
`
`Total
`
`Filterable
`
`Single dose
`Curve of
`plasma levels
`
`70-IOOmg/mz
`Triexponential Monoexponential
`Bi exponential Monoexponential
`Biexponential
`Biexponential
`
`150-500 mg/m2
`Biexponential
`Bi exponential
`
`Biexponential
`
`Initial t 112
`
`Terminal t112
`
`20min
`17.5min
`23min
`>24h
`30.5 h
`67h
`
`20-30 min
`23.6min
`8-10 min
`
`40-45 min
`
`% Urinary
`excretion of
`dose
`
`45%in.48h
`23%in24h
`
`5.4min
`87min
`
`131 min
`354min
`
`98min
`
`6.7->24h
`
`66%in 24h
`65%in24h
`
`CHIP:
`
`Total
`
`Filterable
`
`20-350 mg/m2
`Biexponential Monoexponential
`at low doses,
`biexponential at
`high doses•
`1.75 h (low doses)
`1.08 h (high doses)
`
`0.96 h
`
`64 h
`
`32.3 h (high doses)
`
`15-6lo/oin24h -
`
`however, less toxicity was reported for patients treated
`with the CBDCA-containing combination. The inves(cid:173)
`tigators concluded improved efficacy for the CBDCA
`combination in these patients. Further discussion of these
`plans is presented in Statistical and other considerations
`(below).
`
`B Cisplatin-sensitive diseases, where standard therapy in
`advanced disease has a major palliative effect; (examples
`include small-cell lung, urinary bladder, uterine cervical,
`and head and neck carcinomas. Analog development for
`this group of diseases is the most difficult. One approach
`advocates a traditional phase II trial for this group of
`patients to establish a minimal level of activity before
`proceeding with major phase III trial(s), especially if hints
`of antitumor activity are seen in patients with these tumors
`in phase I trials. Although the determination of this "min(cid:173)
`imal level" requires 40 patients or more, many inves(cid:173)
`tigators feel that this is an ethical necessity before proceed(cid:173)
`ing to a phase III trial requiring 300- 700 patients, where
`most patients achieve palliation rather than cure of their
`disease.
`In practice, the developmental plans have proven to be
`different for each tumor type in this disease category. For
`small-cell lung carcinoma, traditional phase II trials were
`used, followed by a phase III trial comparing one analog
`
`against the parent compound. For urinary bladder and
`uterine cervical carcinomas, randomized phase II trials of
`the analogs were pursued with the endpoint of determin(cid:173)
`ing whether either analog was significantly superior to the
`other, thereby enabling the exclusion of a major difference
`in antitumor activity with acceptable confidence limits.
`Thus, only one analog need be investigated in comparative
`phase III trials against the parent compound with an ac(cid:173)
`ceptable risk of not having picked the "wrong" analog for
`comparison. For head and neck cancer, one investigator
`chose to undertake a randomized addition-type phase III
`trial with treatment arms of methotrexate vs methotrexate
`plus one analog (CBDCA).
`A review of the trial results for each disease is beyond
`the scope of this report. However, published results from
`the use of these analogs in the treatment of patients with
`advanced carcinoma of the uterine cervix provide a
`general illustration for this group. The study began as a
`randomized phase II trial of the analogs [2, 38] and was
`continued to phase III endpoints [39] when an objective
`response rate (complete plus partial responses) of >20%
`was observed in both arms at the completion phase II. At
`the end of the phase III trial, the investigators concluded
`that neither analog appeared to be as active as the parent
`compound against squamous carcinoma of the uterine
`. cervix [39].
`
`Table 7. Advanced disease sensitivity to cisplatin-contruning regi(cid:173)
`mens
`
`Sensitivity
`
`Examples
`
`A. Sensitive and curative
`
`B. Sensitive with major
`palliative effect
`
`C. Resistant or no major
`palliative effect
`
`Germ-cell tumors
`Epithelial ovarian carcinoma
`Small-cell lung carcinoma
`Bladder carcinoma
`Head and neck carcinoma
`Uterine cervical carcinoma
`Non-small-cell lung carcinoma
`Colorectal carcinoma
`Breast carcinoma
`Melanomas
`
`C Diseases that are often not palliated by treatment with
`cisplatin-containing regimens in advanced disease include
`melanoma and colorectal, breast, and non-small-cell lung
`carcinomas. The developmental plans for this category
`were the most straightforward. A traditional phase II trial
`of each analog in minimally pretreated patients was
`planned to determine antitumor efficacy. However, in dis(cid:173)
`eases where many patients were available for clinical tri(cid:173)
`als, randomized phase II trials were carried out in an at(cid:173)
`tempt to select the more active agent before entering
`definitive phase III trials. If the analog demonstrated ef(cid:173)
`ficacy in the initial 14-20 patients entered, an estimate of
`the level of activity in 40-80 patients was determined.
`Published results from a randomized phase II in
`patients with non-small-cell carcinoma of the lung showed
`
`NOVARTIS EXHIBIT 2083
`Breckenridge v. Novartis, IPR 2017-01592
`Page 6 of 10
`
`

`

`Initial Studies (Randomized Phase II Trials)
`
`Definitive Studies (Randomized Phase Ill Trials)
`
`401
`
`A
`
`B
`
`_/CBDCA
`
`/
`/ '
`
`1. One ~nalog appears
`superior
`
`}
`
`®---CHIP
`
`- - - - 2. Neither analog
`appears superior
`
`- - - CBDCA I
`
`@--..._ Cisplatin
`
`1. Analog in each trial
`appears superior to
`cisplatin
`
`- - - CHIP
`
`®-- Cisplatin
`
`2. An analog appears
`superior in one trial
`and cisplatin appears
`superior in the other
`trial
`
`3. Cisplatin appears
`superior in each trial
`
`-®
`
`___.- Best analog
`
`- - Cisplatin
`
`/CBDCA
`
`OR
`
`®-CHIP
`
`'---,... Cisplatin
`
`OR
`
`/
`
`CBDCA
`
`@-CHIP
`
`~ Cisplatin
`
`-®
`
`Either
`_,.,,,,,, analog
`
`- - Cisplatin
`
`--CBDCA
`
`- - - - CHIP
`
`-®
`- ® - - Best analog
`
`---- Cisplatin
`
`Accrue adequate numbers
`to each arm to yield a
`definitive answer
`
`® = Randomize
`Fig. 2. A scheme of possible stepwise clinical evaluations of two cisplatin analogs
`
`the activity of both analogs to have overlapping 95% con(cid:173)
`fidence limits (CBDCA, 7%-25%; CHIP, 1%-13%) with
`similar survival [28]. Although the activity was modest, the
`investigators recommended further studies of CBDCA as a
`component of combination chemotherapy.
`
`Statistical and other considerations
`For the cisplatin-responsive diseases (curative and major
`palliative effect), the optimal trial design to answer the
`questions of relative therapeutic index is a three-way, ran(cid:173)
`domized phase III trial using these compounds at optimal
`doses. The trial must be designed to detect a significant
`decrease in toxicity and, more importantly, to ensure that
`no significant decrease in survival occurs. For ovarian
`cancer, where the 2-year survival is roughly 60%, this
`would require 160 patients per arm to ensure that a
`decrease to 45% 2-year survival would be detected with
`type I and II error limits of 0.10 [ 16). For testicular cancer,
`where the curability in advanced disease is >80%, the
`problem is even more difficult, because one must ensure
`the detection of a 10% decrease in this rate using the same
`type I and II error limits of 0.10. Testicular cancer is not a
`common malignancy, and the three-arm, randomized trial
`would require 265 patients per arm. Obviously, the three(cid:173)
`arm, randomized trial is not practical for all cisplatin-sen(cid:173)
`sitive diseases. An alternative approach enables the ac(cid:173)
`complishment of indirect comparisons in a two-stage
`fashion (Fig. 2, section A). The initial part of the evalua(cid:173)
`tion consists of randomized phase II studies with CBDCA
`and CHIP in selected cisplatin-sensitive (Table 7, category
`B) and -insensitive diseases (Table 7, category C). If one
`analog exhibited a clear advantage in the initial studies,
`
`that analog would be used in the definitive phase III
`studies and the other could be eliminated from further tri(cid:173)
`als in those disease sites.
`The number of patients required for each two-arm,
`randomized phase II study is dependent on the anticipated
`activity of the compounds. The diseases in which a higher
`response rate is anticipated require higher patient numbers
`per arm to ensure a 90% probability of detecting a 15% dif(cid:173)
`ference in the rates. Where response rates are 15%-20%,
`25-30 patients per arm are required, and where they are
`;..,30%, 35-40 patients are required [56].
`A different set of studies are shown in section B of
`Fig. 2 to illustrate another alternative stepwise approach to
`determine the relative therapeutic indices. This approach
`uses two initial-stage studies to plan the definitive study.
`There are three possible outcomes to the initial studies. If
`both analogs are superior to cisplatin, the definitive study
`will involve only the analogs. If one analog is superior in
`one study but cisplatin is superior in the other, the lesser
`analog could be eliminated from the definitive study if the
`apparent difference in the therapeutic indices is significant
`and the test conditions in the two initial studies are the
`same. If neither analog is superior, sufficient numbers of
`patients must be accrued to each arm of the initial studies
`to yield a definitive answer.
`In the stepwise evaluations of CBDCA and CHIP
`planned by the NCI, the two-stage design (section A) was
`used whenever possible because (a) the preclinical results
`showed more similarities between the two analogs than be(cid:173)
`tween either analog and cisplatin; and (b) the section-A
`design required one rather than two studies at the phase II
`level, thus simplifying the planning. The notable exception
`to using this design was the planning of a three-way, ran-
`
`NOVARTIS EXHIBIT 2083
`Breckenridge v. Novartis, IPR 2017-01592
`Page 7 of 10
`
`

`

`402
`
`domized definitive trial in previously untreated patients
`with advanced ovarian cancer after the completion of the
`phase I CBDCA and CHIP trials. It was felt that phase II
`trial results were not needed to plan the definitive trial in
`this disease site because of the number of ovarian cancer
`patients who had been treated in phase I trials and had
`shown a major response.(complete or partial).
`
`Conclusion
`The broad spectrum of antitumor responsiveness to cis(cid:173)
`platin and the high incidence of severe gastrointestinal,
`renal, and neurologic toxicity have spurred the develop(cid:173)
`ment of a number of first-generation cisplatin analogs. The
`majority of these analogs are characterized by: (a) a
`preclinical antitumor spectrum similar to that of cisplatin,
`(b) quantitatively less preclinical renal and gastrointestinal
`toxicity and (c) quantitatively more bone marrow toxicity
`than cisplatin. CBDCA and CHIP are two such analogs
`currently undergoing clinical trials for determination of
`the antitumor and toxicity spectrum of each relative to the
`other and, ultimately, to cisplatin. Because of differences
`in their mechanism of action and antitumor activity, both
`compounds underwent phase I testing. where objective
`responses were reported from trials with each compound.
`Sufficient patient numbers for a particular disease, ethical
`considerations when a palliative effect rather than a cure
`was expected, and statistical considerations were factors in
`forming the proposed approach to the clinical develop(cid:173)
`ment of each analog. The proposed developmental strat(cid:173)
`egy incorporated trial designs based on disease respon(cid:173)
`siveness to cisplatin for assessment of relative disease(cid:173)
`specific antitumor activity. In particular, the toxicity
`results from comparative phase III trials using the
`analog(s) and cisplatin at their maximal doses (alone or in
`fixed-dose drug combinations) will enable assessments of
`relative tolerance. Using the combined results for relative
`disease-specific antitumor activity and relative tolerance,
`relative therapeutic index may be assessed. Using the
`methodology outlined here, we feel that the best analog for
`each particular disease may be determined, providing a
`firm foundation for disease-specific combination proto(cid:173)
`cols and future cisplatin analog development.
`
`Acknowledgements. The authors would like to thank Mr. Wayne P.
`Henry and Ms Sandi Rife for their expert assistance in the
`preparation of this manuscript and Dr. Silvia Marsoni for helpful
`discussions during the early stages of formulating the develop(cid:173)
`mental plans.
`
`References
`I. Alberts D, Green S, Hannigan E, O'Tools R, Mason-Liddil
`N, Surwit E, Stock-Novack D, Goldberg R, Malviya V, Nah(cid:173)
`has W (1989) Improved efficacy of carb