Molecular biomarkers in clinical development: what ... - Future Medicine

Research Article For reprint orders, please contact: [email protected]

Molecular biomarkers in clinical development: what could we learn from the clinical trial registry?

Aim: Objective of this research is to assess whether the trend of stratified medicine widely discussed in scientific literature is translated into real clinical trials registered in ClinicalTrials.gov. Methods: By semi-automatic screening of over 150,000 trials, we filtered trials with stratified biomarker to analyze their therapeutic focus, major drivers and elucidated the impact of stratified biomarker programs on trial duration and completion. Results: >5% of trials are using molecular biomarker for stratification; duration of such trials is longer. 21% of them are done in late stages and Oncology is the major focus. Conclusion: Trials with stratified biomarker in drug development has quadrupled in last decade but represents a small part of all interventional trials reflecting multiple co-developmental challenges of therapeutic compounds and companion diagnostics. Keywords: cardiovascular disorders • digestive system disorder • hematologic disorders • infectious diseases • MS-based proteomic profiling • metabolic diseases • neurological diseases • oncology • respiratory disorders • stratified molecular biomarker

Background Since the beginning of this century, the crisis of the pharmaceutical industry has been reflected by unsustainably high expenditures and low output associated with a high drug failure rates [1–3] . One of the major causes of expensive drug failure is the marginal disease improvement compared with current standard of care when analyzed in large population of late stage trials [4] . This drove one of the paradigm shifts of pharmaceutical drug discovery to move the focus from blockbusters to nichebusters, for example, therapies targeted towards specific target molecules of specific patient populations coined as targeted therapy. The success of targeted therapy depends on identifying stratified biomarker in patient before therapeutic intervention. Biomarker is a characteristic that objectively measured and evaluated as an indicator of normal biological processes or pharmacologic responses to a therapeutic intervention. Whereas stratified biomarker is a biomarker that will stratify the patients

10.2217/PME.14.27 © 2014 Future Medicine Ltd

Avisek Deyati1, Rama Devi Sanam2, Sreenivasa Rao Guggilla2, Vijaya Rao Pidugu2 & Natalia Novac*,3 1 Department of Bioinformatics, Fraunhofer Institute for Algorithms & Scientific Computing (SCAI), Schloss Birlinghoven, 53754 Sankt Augustin, Germany 2 GVK Biosciences Pvt. Ltd., Hyderabad, 500076, India 3 Merck Serono, 250 Frankfurter Strasse, 64293, Darmstadt, Germany *Author for correspondence: Tel.: 00496151726846; Fax: 0049615172916846; natalia.novac@ merckgroup.com

prior to the treatment which will likely to be benefited from the therapy [5] . The success stories of targeted therapy can be exemplified with commercialization of trastuzumab, cetuximab, imatinib, gefitinib. The trend is continuously rising with the most recent examples of Vemurafenib approved with the companion genetic test for the BRAF mutation for late stage melanoma and Crizotinib approved in combination with the companion genetic test for the ALK gene for late stage lung cancer [5–8] . Despite the obvious advantages of the stratified medicine approach for patients and payers, such as prevention of overtreatment and an early decision for an alternative therapy [4] , the business incentive for pharmaceutical companies to invest into the co-development of the stratified molecular biomarker early on is less clear [9] . Also financially it can be seen as a burden for the pharmaceutical companies as they have to bear the added cost of companion diagnostic development and yet have to cope with the reduced mar-

Personalized Medicine (2014) 11(4), 381–393

part of

ISSN 1741-0541

381

Research Article Deyati, Sanam, Guggilla,

Pidugu & Novac

ket size due to patient stratification [9] . However recent observations show that stratified biomarker-aided drug discovery is commercially more viable for pharmaceutical companies than post-approval research for diagnostic testing. For example clinical development of trastuzumab and imatinib with stratified biomarker have enhanced clinical, commercial success but post approval application of KRAS mutation test for cetuximab and panitumumab might not be commercially very rewarding [10] . In order to understand the degree to which biomarker programs are implemented as well as the current trends and risks of inclusion of stratified biomarkers in clinical trials, we decided to analyze more than 150,000 clinical trials entries accumulated in ClinicalTrials.gov database [39] . Since the inception of ClinicalTrials.gov database in 2000 (in response to US congress law obliging NIH to publish private and federally-sponsored trials) it has rapidly become the most favored publicly available search engine in clinical trials registry covering trials from all the geographical locations in the world. Although the database was launched in the year 2000, clinical trials with start date as early as 1970 (NCT00005125) has been incorporated in it, thus representing 43 years of clinical research. Despite the known issues with the updates, consistency and completeness of the data [11,12] , it is the oldest and largest clinical trial registry containing the information on more than 150,000 trials (August 2013). Choosing between the clinical trials registries, we reckoned that the analysis of the largest database will give us a representative picture of the historic and current trends on the use of stratified molecular biomarkers in clinical trials. The major questions we investigated here are: • Q1. How many clinical trials used stratified biomarkers and whether the stratified medicine trend is reflected in the increased number of trials using the approach? • Q2. Who are the major sponsors of biomarker programs in clinical trials? • Q3. What are the major indications where stratified biomarkers are explored? • Q4. Which stage of clinical development the biomarker program is applied in clinical trials? • Q5. What are the technologies used for the biomarker discovery? • Q6. Does the inclusion of stratified biomarkers into clinical trial increase the trial completion time?

382

Personalized Medicine (2014) 11(4)

• Q7. Does the inclusion of stratified biomarkers affect the chances of trial being completed? Methods In addressing above mentioned questions, we downloaded the entire database as XML files on 02 August 2013 for analysis. On that date the total number of clinical trials registered in the database was 150,504. The majority of the studies (121,922) were interventional, for example, analyzing clinical outcome after intervention. The remaining 27,886 studies were observational. The analysis started with the selection of three groups. Interventional trials (Group 1)

Focusing on 121,922 interventional trials, we further filtered trials by excluding those with unknown “Overall status” to remove uncertainty regarding updates of the database. In avoiding time confounders interventional trials with “start date” between 1991 to 2013 were programmatically filtered as first trial with stratified biomarker was registered in 1991 (NCT00001271). In many trials “intervention name” and targeted disease (“condition”) field can be empty, so we programmatically checked “intervention name” and “condition” field of each trial to select only those 60,629 interventional trials with a drug/biologics term as “intervention name” and a disease term as “condition”. Interventional trials with biomarker as outcome measure (Group 2)

In selecting Group 2 the “outcome measures” field of interventional trials was checked and only those with “biomarker” in it, were selected. Next, similar to Group 1 “start date”, “intervention name” and “condition” fields were programmatically checked. Finally we filtered 4745 trials with biomarker as outcome measure. To check the quality of the screening method for the selection of Group 2 trials, we have randomly checked 10% in which could not find out any false positives. Interventional trials with stratified molecular biomarker (Group 3)

The prime focus of our analysis was trials with stratified molecular biomarker and to have a list of such trials, first we developed a set of key words derived from the manual annotation of 80 studies (when searched ClinicalTrials.gov with “Cetuximab AND KRAS”). The aim of this manual annotation was to look for intuitive words which can possibly filter a trial with stratified biomarker and a search keyword was developed (Figure 1) . 22,273 trials were filtered by search-

future science group


Work flow for selection of trials without stratified biomarker

Exclude unknown | Interventional studies trials | Start date: 1991 to 2013 | Contains name of drug/biologics and disease

ClinicalTrials.gov

Yes

Work flow for selection of trials with stratified molecular biomarker Searched with “Cetuximab and KRAS”

ClinicalTrials.gov

Biomarker as outcome measure?

Group 1: Interventional trials (60,629)

Manually curate all 80 studies to find out list of key words to find out trials with stratified biomarker

80 Studies

Group 3: 1,701 trials with molecular stratified biomarker

Yes

Research Article

Group 2: Interventional Trials with Biomarker as outcome measure (4,745)

(biomarker OR resistance OR marker OR positive OR mutation OR miRNA OR sensitivity OR sensitive OR genomic OR microarray OR proteomic OR metabolomic OR polymorphism OR SNP OR Negative) NOT (respirat OR airway) | Exclude Unknown | Interventional Studies trials

Automatic tagging NCBI genes | Drugs/Biologics

Manual 5,420 trials curation

22,273 trials

Work flow for the meta analysis Perl program (based on DDT of ClinicalTrials.gov) Q1

Q2 Genetic

Q4

Q3

Q5

Yes

Manual curation of the disease term to segment into major therapeutic category

Q7

Q6

Yes

Completed?

Completed

Yes Terminated

(Start date – Completion date) Diseases & aliases mapped to Mesh Disease IDs

Completion time (in months) comparison between Group 2 and Group 3 for trials targeting six therapeutic category

% of trials comparison, as completed, terminated between Group 1 and Group 3 for trials targeting six therapeutic category

Figure 1. Workflow for the selection of trials with, without stratified molecular biomarker and meta analysis of the ClinicalTrials.gov [39] . Q1, Q2, Q3, Q4, Q5, Q6, Q7: major investigated questions as mentioned in introduction.

ing the ClinicalTrials.gov with the developed keyword. Focusing on the trials that are using molecular biomarkers we further filtered the resulted list by excluding those trials without any gene or protein names by automated tagging reduced the list to 5420 interventional studies. All the 22,273 trials with outcome of the above mentioned tagging and presence of “intervention name” are presented in Supplementary Table 1 (see Supplementary data online at: www.futuremedicine.com/doi/full/10.2217/PME.14.27). Automated tagging was done by running an internally-developed script containing a list of all gene and its synonyms obtained from NCBI database [40] . Using the script, we screened the important xml fields of each clinical trial, for example, “title”, “purpose”, “official title”, “primary outcome measures”, “secondary outcome measures”, “detailed description” and “keywords” to filter out the trials with gene names. Further focusing on the trials


that use molecular biomarkers for patient stratification prior to treatment; we manually curated those 5420 trials. Finally 1701 trials with stratified molecular biomarkers were filtered and these trials were further analyzed and became the basis of this paper. In ensuring the authenticity of this screening, we randomly checked 10% of the trials that did not appear in the result set (i.e., 5420–1701) for the presence of any gene names. We found out that none of these trials were false positive either. The examples of the trials filtered out by manual curation as trials with stratified molecular biomarker, please refer to Supplementary Table 2. Trials in Group 1, Group 2 and Group 3 are not overlapping. Identification of disease terms & segmentation into therapeutic categories

One of the major aims of this analysis was to elucidate therapeutic focus of Group 3 trials. To achieve

www.futuremedicine.com

383


Pidugu & Novac

it, at first “condition,” for exmaple, targeted disease of 1701 trials belonging to Group 3 was collected. Next, all those targeted diseases were manually curated and segmented into 15 majortherapeutic areas based on MeSH disease tree. Next important focus of the analysis was to investigate the impact of stratified biomarker on trial duration of “completed” trials and trial’s status across different therapeutic categories curated in the earlier step. In order to have a balanced comparison of trial duration between target group (Group 3) and control group (Group 2), the duration period were calculated in months. The assessment of trials status was determined as follows: a trial was considered successful when “Overall status” was “completed” whereas, a trial was considered unsuccessful if “Overall status” was “terminated” (definition of terminology [41]). The comparison has been done between Group 3 and Group 1 i.e. all other interventional trials excluding Group 3 (Figure 1) . For the quantitative analysis, formatting has been completed with Perl programming language based on the Document Type Definition (DTD) [42] of ClinicalTrials.gov. To map diseases related synonyms to a unique identifier, MeSH disease dictionary was applied. Figure 1 shows a flow chart describing the logic and steps of our meta analysis. The method has been described in detail in the Supplementary material called “Detailed Description.txt”. All the manually curated individual disease indications belonging into 15 major therapeutic categories are listed in Supplementary Table 3. Results How many clinical trials use molecular biomarkers for patients’ stratification?

Our primary goal was to understand whether the stratified medicine trend as a consequence of the genetic revolution, widely discussed in biomedical scientific forums; is actually translating in the use of molecular biomarkers in clinical trials. In investigating that at very first step we collected “start date” of each trial and calculated the percentage of clinical trials that are using molecular biomarkers for patients’ stratification. Comparing them to the total number of clinical trials registered in the database we found 1.39% of all interventional clinical trials (i.e. 1,701 trials out of total 121,922 interventional trials) belong to this category. In order to figure out what is the trend in implementation of biomarkers in clinical research, we analyzed the historical record of trials with stratified molecular biomarker (TSMB). The first trial using stratified molecular biomarker was registered in 1991 when CD22(+) B cell lymphoma patients were selected for treatment with IgG-RFB4-SMPT-dgA antibodies (NCT00001271).

384


Since then, the number of such clinical trials is steadily growing and attending the peak in 2011 with 214 trials registered that year (Figure 2A) . We further calculated year- wise proportion of trials with stratified molecular biomarker compare to total number of interventional trials with 95% confidence interval (CI) starting from 2000 to August, 2013. In Figure 2B, we have plotted year wise lower and upper limits of the proportion with 95% CI. Based on Figure 2B with 95% CI we can see that less than 5% of all interventional trials are using molecular biomarker for patient stratification. All the trials with stratified molecular biomarker and its start year can be found in Supplementary Table 4. What are the key technologies for biomarker identification used in clinical trials?

In order to understand the technologies being applied in clinic for the detection of stratified biomarkers and their frequency of application, the “intervention name” from each TSMB was selected when “intervention type” is “Genetic”. Unfortunately less than 10% of trials specify this information in the registry. TSMB trials with specified technologies are listed in Supplementary Table 5. After programmatically retrieved the technologies, they were manually curated and then segmented into three different omics technologies, for example, genomics, transcriptomics and proteomics. Genomic technologies including detection of mutations and polymorphisms by PCR, gene sequencing and cytogenetic analyses are used in 50% of trials (Supplementary Figure1) . Genomic technologies are followed by transcriptomics analysis combining various technologies of gene expression arrays and used in about 40% of trials reporting the biomarker detection techniques. Different to the traditional single biochemical and histopatho-logical measurements, expression profiles represent a fingerprint containing multiple biomarkers, which collectively indicate a particular pathophysiology [13] . The rest 10% of the trials are using proteomics technologies for detection of stratified biomarker, starting from Western Blotting in early studies and ending with mass spectrometry based proteomic profiling (e.g., NCT01658566, NCT00601913). The leading role of genomic technologies in biomarker detection clearly represents the current trend in clinical biomarker discovery reflecting both stability of the genomic signal and commoditization of the genomic technologies. Who are the major funding organizations in the clinical biomarker field?

Focusing on the 1,701 trials, we extracted the “lead sponsor” of the trials; to investigate the major players in the field investing heavily into the stratified bio-



B

A

Research Article

Lower limit with 95% CI Upper limit with 95% CI 0.045

Proportion of Group 3 trials to all interventional trials

200

150 100

50

0

0.040 0.035 0.030 0.025 0.020 0.015 0.010 0.005

Start year of trials with molecular stratified biomarker

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

2003

2002

2001

1991 1994 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013

0 2000

No. of trials with stratified biomarker

250

Start year

Figure 2. Growth and year-wise proportion of trials with stratified molecular biomarker.

marker programs. Three pharmaceutical companies such GSK, Roche and Novartis appear to be the main industrial sponsors with 3.7, 3.1 and 2.7% (Figure 3) of all trials with stratified molecular biomarker. Therefore it is not surprising that those companies are behind most recent breakthroughs in the field of stratified medicine. GSK compound Dabrafenib with the companion diagnostics for detection of BRAF mutation has received US FDA approval for the treatment of melanoma [14] . Another GSK’s MEK inhibitor trametinib has been approved with BRAF mutation as stratified biomarker [15] . Roche antibody–drug conjugate trastuzumabemtansine has been approved for the treatment of Her2-positive advanced breast cancer patient [16] . Another Roche compound vemurafenib has been approved for the treatment of BRAF-mutated metastatic melanoma [17] . Imatinib of Novartis has got the approval for the treatment of leukemia patients with specific PDGFR and C-Kit mutations [18] . Pfizer compounds Crizotinib with the companion diagnostic (ALK5 mutation) [19] and Maraviroc with the companion diagnostic TM test for the viral tropism have also been approved [20] . Vast experience of these companies in conducting trials with stratified biomarker allowed for these targeted approaches and in case of Crizotinib unprecedentedly shortened the development of the drug leading to millions of savings for the company [7] . National Cancer Institute (12.35%) and NIAID (4.41%) seem to have the major academic drivers of tri-


als with stratified molecular biomarker. As European agencies are not obliged to fill into the database, we cannot see any European academic player in the top 15 clinical research organizations. Other proprietary databases such TrialTrove or PharmaProjects would be better for the analysis of the European situation in the field of clinical biomarker research. Trials with stratified molecular biomarker and its lead sponsor can be found in Supplementary Table 6. What are the major indications where biomarkers are searched for?

According to our analysis (Figure 4), oncology represents more than 75% of all the trials with stratified biomarker program. Infectious disorders are next major focus of trials with stratified biomarker representing about 10% of all the analyzed studies (Figure 4A) . Other therapeutic areas combined, represent the rest 15% of the trials with molecular stratified biomarker and metabolic diseases leading this group (Figure 4B). Closer look at the individual indications of oncology therapeutic area reveals that most of the registered stratified molecular biomarker studies are in the field of breast cancer comprising 478 studies and constituting 28% of all trials (Figure 4C) . This extensive clinical effort is translated into the marketed companion diagnostics and breast cancer patients are benefiting from it. In this indication hormone-dependency tests, genetic susceptibility tests (such as BRCA1/2 mutations) and gene expres-


385


Pidugu & Novac

12 10 8 6 4 2

University of Washington

Dana-Farber cancer institute

Boehringer Ingelheim pharmaceuticals

Massachusetts general hospital

Sanofi

AstraZeneca

M.D. Anderson cancer center

Bristol-Myers Squibb

Memorial Sloan-Kettering cancer center

Pfizer

Novartis pharmaceuticals

Hoffmann-La Roche

GlaxoSmithKline

NIAID

0

National Cancer Institute (NCI)

% of trials with stratified biomarker

14

Top 15 funding organisation for trials with stratified molecular biomarker Figure 3. Major funding organizations sponsoring trials with stratified molecular biomarker.

sions analysis (e.g., Mammaprint and OncotypeDX) became a part of a standard clinical care [21,22] . Lung cancer is the second most frequently targeted (20.8%) by trials with stratified molecular biomarker with 353 studies registered in ClinicalTrials.gov. Years of clinical research are translated in a number of approved biomarkers in this therapeutic area, such as Crizotinib approved in combination with the companion genetic test for the ALK5 gene for late stage lung cancer [7] . Leukemia and Lymphoma are the next largest oncological indications benefiting from the early discovery of stratified molecular biomarkers such as Philadelphia chromosome or other genetic translocations such as RAR-PML fusions that brought to the early discovery of the target-specific treatments, celebrating the first wins of molecular biology in clinics [22] . Trials with stratified molecular biomarker and its targeted disease indications can be found in Supplementary Table 7. At which phase molecular biomarkers are used for stratification?

Further analysing the stage of the clinical development at which the stratification biomarkers are explored,

386


we extracted the information on the “phase” from all 1,701 TSMB (Figure 5) . Approximately 93% of trials with stratified molecular biomarker contain the information on clinical phase. According to our analysis most of the trials, around 60%, are in phase II. This category includes the trials indicated to be in the phase I-II as well. 15% trials with stratified molecular biomarker are in phase I and III. It was interesting to see that almost 6% of all stratified molecular biomarkerassociated trials are conducted at phase IV. It clearly reflects that search for stratified biomarkers are continued in the post-marketing phase as well. Trials with stratified molecular biomarker and its clinical phase can be found in Supplementary Table 8. Does the biomarker program affect clinical trial duration and its chance to be completed?

Our next question was to analyse whether the inclusion of stratified molecular biomarker has any impact on trial duration. To answer that clinical trial duration was calculated in months as (“start date” - “completion date”). Statistical distribution of clinical trials duration, targeting disease indications of Group 2



A

Research Article

259 191 1317

Oncology

Infectious diseases

Other

Other targeted therapeutic categories by trials with molecular stratified biomarker

5 Peritoneal neoplasms

Ovarian neoplasms

Adenocarcinoma, mucinous

0 Melanoma

Genital disorders

Vision disorders

Auto immune

Skin diseases

Urological diseases

Musculoskeletal diseases

Endocrine system diseases

Neurological diseases

Respiratory disorders

Cardiovascular diseases

Digestive system disorders

Hematologic disorders

0

10

Prostatic neoplasms

5

15

Colorectal neoplasms

10

20

Lymphoma

15

Leukemia

20

25

Lung neoplasms

25

30

Breast neoplasms

30

% of trials with stratified biomarkers

C

35

Metabolic diseases

% of trials with stratified biomarkers

B

Major types of oncology indications targetted by trials with molecular stratified biomarker

Figure 4. Major targeted therapeutic areas by trials with stratified molecular biomarker. (A) Major indication areas targeted by trials with stratified biomarker; (B) Other targeted indication areas; (C) Major types of targeted cancer.

(interventional trials with biomarker as outcome measure) and Group 3 (interventional trials with stratified biomarker) and falling into six therapeutic areas were compared in Figure 6A . As can be seen in Figure 6A , the addition of molecular stratified biomarkers into clinical trials targeting oncology, infectious diseases and neurologic disorders extend the trial duration, evident from the difference in the central tendency of data (i.e.. median). We also calculated therapeutic area specific mean trial duration and difference of means with 95% CI between Group 2 and Group 3 trials by Welch two sample t-test and represented in Table 1. With 95% CI stratification step increases trial duration by 13.3–2.7 months in oncology, by 27–7.5months in infectious diseases and by 32.4–8 months in neurologic disorders. In metabolic and cardiovascular diseases the application of stratified biomarker seems to be less significant in affecting clinical trial duration. On the other hand in respiratory disorders stratified biomarkers on mean scale shorten the trial duration by 19.7–8 months. All the completed trials belonging


to Group 2 (WB) and Group 3 (SB) and falling into above mentioned six therapeutic categories are presented in Supplementary Table 9 along with its “start date”, “completion date”, trial duration in months. As we see, the duration of trials could be affected by the inclusion of stratified molecular biomarker; however question arises if completion itself is affected by inclusion of stratified molecular biomarkers? In addressing this question we compared the proportion of “Completed” trials across six most targeted therapeutic areas between Group 1 (all interventional trials) and Group 3 (interventional trials with stratified biomarker) as a measure of success. To measure the rate unsuccessful trials proportion, “Terminated” trials across same six targeted therapeutic areas between Group 1 and Group 3 was compared. Proportion of “Completed” trials across six therapeutic categories of each group was calculated based on number of “Completed” trials in given therapeutic category of a group divided by total number of trials targeting the same therapeutic category of the same group. The proportion of “Ter-


387


Pidugu & Novac

% of trials with stratified biomarker

70 60 50 40 30 20 10 0

Phase 0

Phase 1

Phase 2

Phase 3

Phase 4

Clinical phases of trials with stratified molecular biomarker Figure 5. Phase wise distribution of trials with molecular stratified biomarker.

minated” trials was calculated using the same formula. Patient stratification by molecular biomarkers do seem to significantly affect the fate of the trial whether it will be successful or unsuccessful (Figure 6B, & C). In metabolic, cardiovascular and neurologic disorder, trial termination rate is significantly higher in Group 1 (trials without a biomarker). However fewer trials targeting respiratory and oncology disorders, are completed in Group 3 compared with the Group1 trials. If in other therapeutic areas the average percentage of terminated studies is approximately 8%, addition of stratified biomarkers to trials targeting respiratory disorders, doubles the chance of the trials to be terminated prematurely. In the case of oncology every third molecular biomarker trial is completed compared with every second in the non-biomarker group. According to our analysis, more than one third of Group 3 trials are started after 2009 and oncology being most frequently targeted (Figure 2 ; 4A). Knowing that the median trial duration is ∼45 months (Figure 6A), most of them are still ongoing and ∼31% (411 out of 1317) of oncology trials with stratified biomarker are still in the recruiting phase. At the same time slightly higher termination of stratified biomarker trials targeting oncology, reflects the hard road of clinical development (well standardized tests, trained personnel and dedicated resources) targeting highly heterogeneous diseases like cancer [23,24] . Among all studied therapeutic areas respiratory disorders are more challenging for stratification due to heterogeneity of the population and the fact that biomarker studies in respiratory disorders are still in their infancy [25–28] . The result is evident in our study, which shows that more clinical

388


trials are terminated and less completed when stratified compared with the non-stratified group. Group 3 and Group 1 trials with its ‘overall status’ can be found in Supplementary Table 10 and Supplementary Table 11. Discussion FDA in its critical path initiative emphasized on applying biomarkers as an essential tool to combat current situation of pharmaceutical industries suffering from late stage failures and lack of successful pipeline portfolios. Number of recent publications suggested that more efficient model of pharmaceutical pipeline can be designed by applying biomarkers in all stages of drug discovery and development i.e. emphasizing biomarker usage from target identification to drug marketing [29,30] . However, until now the usage of patient stratification biomarker in late stage clinical trials stands out among other biomarker applications to cope with the most alarming issue of expensive late stage failures. According to the simulation-based analysis performed by the FDA and MIT consortium, early biomarker program is predicted to be an economically valuable model for pharmaceutical development [31] . Corroborating this prediction, Parker et al. have shown that the application of Her2 in the development of anti breast cancer treatment reduced the clinical trial risk by 50% and lead to 27% cost savings [32] . On a global scale, it has been quantitatively demonstrated that probability of phase to phase transition is 15–19% higher for anti-cancer treatments if their trials include biomarker programs compared with those without [33] . All these findings indicate that early application of stratified molecular biomarker



Research Article

A

Trial duration in months

150

100

50

0 0WB

0SB

IWB

ISB

MWB

MSB

RWB

RSB

CWB

CSB

NWB

NSB

Trials without biomarker (WB) and with stratified biomarker (SB) targeting six therapeutic areas

Completed trials targeting six therapeutic areas

Terminated trials with stratified biomarker Terminated trials without stratified biomarker

Respiretory disorders

Neurologic disorders

Cardiovascular disorders

Metabolic disorders

Infectious diseases

20.00 18.00 16.00 14.00 12.00 10.00 8.00 6.00 4.00 2.00 0.00 Oncology

% of trials Respiratory disorders


Cardiovascular disorders

Metabolic disorders

Infectious diseases

80.00 70.00 60.00 50.00 40.00 30.00 20.00 10.00 0.00

C

Completed trials with stratified biomarker Completed trials without stratified biomarker

Oncology

% of trials

B

Terminated trials targeting six therapeutic areas

Figure 6. (A) Comparative statistical distribution of clinical trials duration between Group 2 and Group 3. (B) Comparative proportion of completed trials between Group 1 and Group 3. (C) Comparative proportion of terminated trials between Group 1 and Group 3. Impact of stratified biomarker program on trial duration and completion. The three-letter abbreviations used along the x axis in (A) represents each category. The first letter of the abbreviation represents targeted therapy (O: Oncology; I: Infectious diseases; M: Metabolic disorders; R: Respiratory disorders; C: Cardiovascular diseases; N: Neurologic disorders) followed by two other letters representing groups (WB: Group 2; SB: Group 3)

may significantly improve the success of clinical development. As representative of worldwide clinical trials this analysis shows that 21% of the stratified trials are done in the later stages (Phase III and IV; Figure 5) with significant efforts in post-marketing research. A more comprehensive picture of worldwide clinical tri-


als can be drawn by inclusion of proprietary clinical trial registries such as Trialtrove and PharmaProjects into our analysis; however the proprietary laws do not allow us to publish these results. ClinicalTrials.gov itself has reported issues with the updates, consistency and completeness of the data particularly in those trial


389


Pidugu & Novac

Table 1. Therapeutic area specific mean trial duration and difference of means with 95% CI between Group 2 and Group 3 Therapeutic area

Group 2 abbreviation, Mean Group 3 abbreviation, Mean Difference of means between trial duration in months trial duration in months Group 2 to Group 3 with 95% CI

Oncology

OWB, 42.2

OSB, 50.2

-13.3 to -2.7

Infectious diseases

IWB, 32

ISB, 49.3

-27 to -7.5

Metabolic disorders

MWB, 26.6

MSB, 27.5

-8 to 6.2

Respiratory disorders

RWB, 25

RSB, 11.2

8 to 19.7

Cardiovascular disorders CWB, 27.8

CSB, 30.2

-18 to 13.3


NCB, 41.2

-32.4 to -8

NWB, 21

files collected before the database launch [11,12] . Surely a complete manual curation of all 150,000 trial registry files may enhance the accuracy of our analysis however such an effort is immensely time consuming involving multiple scientific annotators, which is beyond the scope of this research. Thus our semiautomatic approach in identifying trials with stratified biomarker from the oldest and largest clinical trial registry and successive analysis is a representative study of worldwide trend in clinical trials. The analysis also shows with 95% CI, that less than 5% of all interventional trials are using molecular biomarker for patient stratification. Some reasons for slow adoption of biomarker approaches in the clinics are summarized below. First of all stratification biomarker discovery requires an in depth understanding of disease mechanism and drug mode of action. Demonstration of clear relationship of biomarker changes and disease progression or treatment success requires substantial effort and resources spent early in pre-clinical research. However the worldwide industrial trend to reduce the costs of research and to move faster to the clinics hampers early biomarker exploration and validation. Collaboration with academic research and contract research organizations specializing on the biomarker discovery considered to be the future path for improvement of the situation. Secondly, co-development of a drug along with its diagnostic test requires close alignment of these processes. Such parallel development is often very challenging due to different stages, time-lines and expertise involved [34–38] . This opens opportunity for a prospective collaboration between pharmaceutical and diagnostic companies resulting in continuous rise of such alliances observed in the last 5 years MDX Monitor [36,,43] . Most importantly ever-changing regulatory requirements and lack of uniformity of companion diagnostics regulation in key drug markets create significant financial challenge for stratified medicine approval [24] . As stratified medicine is targeted towards smaller patient

390


population and suffers from lack of uniform regulation between countries, the model becomes less appealing to pharmaceutical organizations. A consensus regulation and early, open communication between pharmaceutical companies along with diagnostics partner and regulatory bodies can be the future solution. Conclusions This analysis of stratified molecular biomarker trials registered in ClinicalTrials.gov database clearly shows that the molecular biomarker trend catches both industrial and academic clinical research with oncology being in the forefront of the personalized medicine. However percentage of the studies including patients’ stratification based on molecular differentiation is still very low (less than 5%) reflecting all the challenges of biomarker development. We believe that this analysis will prompt the clinical researchers to invest more into development of diagnostic tests particularly in the early stages of therapeutic development. The analysis will appeal regulators to develop a more attractive consolidated path for companion diagnostic development encouraging early molecular biomarker development. The success of personalized medicine lies in joint forces of both academic and industrial world supported by aligned regulatory bodies resulting in the benefit for patients, payers and pharmaceutical companies. Future perspective Although our analysis shows that the proportion of clinical trials, that use molecular biomarkers for patients’ stratification; is currently relatively low. But we believe that the situation will dramatically change in the next decades. We see three major drivers that will contribute into future acceleration of the stratified biomarker field. • Great technological advances in the genome sequencing field and extremely rapid commoditization of personal genome sequencing will soon



lead to the situation that the individual genome sequence information will be as common as blood group knowledge today. This information will allow and urge the physicians and drug-makers to create customized treatment approaches; • Patient advocacy groups are gaining more and more visibility and strength these days. Soon the patients will have more control over clinical research directing it towards tailored solutions persuading both drug industries and authorities to make it happen; • Increasing pressure of regulatory authorities on pharmaceutical industry to demonstrate efficacy and differentiation of the new products compared with those on the market is boosting diagnostic field and will result in more clinical research in this area. With all this in mind we think that the number of stratified trials will at least double in the next 10 years, which will ultimately lead to customization of the treatment and higher satisfaction of patients enjoying the truly personalized medicine.

Research Article

Acknowledgements Prior to design the research, we consulted with specialists at ClinicalTrials.gov and incorporated their advice into the applied methodologies.

Financial & competing interests disclosure The authors have no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties. No writing assistance was utilized in the production of this manuscript.

Ethical conduct of research The authors state that they have obtained appropriate institutional review board approval or have followed the principles outlined in the Declaration of Helsinki for all human or animal experimental investigations. In addition, for investigations involving human subjects, informed consent has been obtained from the participants involved.

Summary Points Current trend of trails with stratified biomarker • Less than 5% of current registered clinical trials are using molecular biomarker stratification • Since the inception of trials with stratified molecular biomarker in 1991, this field has witnessed a steady growth reaching 214 trials in 2011.

Late-stage exploration of patient stratification • 21% of trials with stratified biomarkers are implemented in late stages (Phase III and IV)

Major focus of stratified medicine • Oncology is the major focus of trials with stratified biomarker.

Major industrial drivers of stratified medicine • GSK, Roche and Novartis appear to be the main industrial sponsors of biomarker trials contributing to more than 40 trials each.

Leading technology for identification of stratified biomarker • Genomics is the most frequently used technology for the selection of stratified molecular bio-marker.

Impact of stratification step in trial duration • Stratification step in oncology, infectious and neurological diseases increase trial duration.

Future path of stratified medicine • This analysis will prompt the clinical researches to invest more into development of diagnostic tests particularly in the early stages of therapeutic development and will appeal the regulators to develop more attractive consolidated path for companion diagnostic development encouraging early molecular biomarker development.

References

3

Pammolli F, Magazzini L, Riccaboni M. The productivity crisis in pharmaceutical R&D. Nat. Rev. Drug Discov. 10(6), 428–438 (2011).

•

Current challenges of drug discovery and development.

4

Frank R, Hargreaves R. Clinical biomarkers in drug discovery and development. Nat. Rev. Drug Discov. 2(7), 566–580 (2003).

5

Trusheim MR, Berndt ER, Douglas FL. Stratified medicine: strategic and economic implications of combining drugs and clinical biomarkers. Nat. Rev. Drug Discov. 6(4), 287–293 (2007).

Papers of special note have been highlighted as: ▪ of interest; ▪▪ of considerable interest 1

Arrowsmith J. Trial watch: Phase II failures: 2008–2010. Nat. Rev. Drug Discov. 10(5), 328–329 (2011).

•


2

Arrowsmith J. Trial watch: Phase III and submission failures: 2007–2010. Nat. Rev. Drug Discov. 10(2), 87 (2011).

•




391

Research Article Deyati, Sanam, Guggilla, ••

Impact of stratified biomarker for targeted therapy.

21

6

Dean E, Lorigan P. Advances in the management of melanoma: targeted therapy, immunotherapy and future directions. Expert Rev. Anticancer Ther. 12(11), 1437–1448 (2012).

Tessari A, Palmieri D, Di Cosimo S. Overview of diagnostic/ targeted treatment combinations in personalized medicine for breast cancer patients. Pharmgenomics Pers. Med. 7, 1–19 (2013).

22

7

Ou SH, Bartlett CH, Mino-Kenudson M, Cui J, Iafrate AJ. Crizotinib for the treatment of ALK-rearranged non-small cell lung cancer: a success story to usher in the second decade of molecular targeted therapy in oncology. Oncologist 17(11), 1351–1375 (2012).

Deyati A, Younesi E, Hofmann-Apitius M, Novac N. Challenges and opportunities for oncology biomarker discovery. Drug Discov. Today 18(13–14), 614–624 (2013).

23

Ludwig WD. [Possibilities and limitations of stratified medicine based on biomarkers and targeted therapies in oncology]. Z Evid Fortbild Qual Gesundhwes 106(1), 11–22 (2012).

24

Deyati A, Younesi E, Hofmann-Apitius M, Novac N. Challenges and opportunities for oncology biomarker discovery. Drug Discov. Today 18(13–14), 614–624 (2013).

25

Taylor DR. Using biomarkers in the assessment of airways disease. J. Allergy Clin. Immunol. 128(5), 927–934 (2011).

26

Ding QL, Sun SF, Cao C, Deng ZC. Association between angiotensin-converting enzyme I/D polymorphism and asthma risk: a meta-analysis involving 11,897 subjects. J. Asthma 49(6), 557–562 (2012).

27

Lee YH, Choi SJ, Ji JD, Song GG. The CTLA-4 +49 A/G and -318 C/T polymorphisms and susceptibility to asthma: a meta-analysis. Mol. Biol. Rep. 39(8), 8525–8532 (2012).

28

Penaloza A, Roy PM, Kline J. Risk stratification and treatment strategy of pulmonary embolism. Curr. Opin. Crit. Care 18(4), 318–325 (2012).

8

Ruzzo A, Graziano F, Canestrari E, Magnani M. Molecular predictors of efficacy to anti-EGFR agents in colorectal cancer patients. Curr. Cancer Drug Targets 10(1), 68–79 (2010).

9

Davis JC, Furstenthal L, Desai AA. The microeconomics of personalized medicine: today’s challenge and tomorrow’s promise. Nat. Rev. Drug Discov. 8(4), 279–286 (2009).

••

Challenges and opportunities of targeted therapy.

10

Loupakis F, Vasile E, Santini D, Masi G, Falcone A, Graziano F. EGF-receptor targeting with monoclonal antibodies in colorectal carcinomas: rationale for a pharmacogenomic approach. Pharmacogenomics 9(1), 55–69 (2008).

11

Wadman M. Verdict on clinical trials registries? Good, but must do better. Nat. Rev. Drug Discov. 5(3), 175–176 (2006).

12

Innocenzi A, Obialero M, Zanetti PP, Personnettaz E. Proposal of a new cutaneous incision in radical mastectomy. Minerva Chir. 39(9), 785–790 (1984).

29

Bhattacharya S, Mariani TJ. Array of hope: expression profiling identifies disease biomarkers and mechanism. Biochem. Soc. Trans. 37(Pt 4), 855–862 (2009).

Colburn WA. Biomarkers in drug discovery and development: from target identification through drug marketing. J. Clin. Pharmacol. 43(4), 329–341 (2003).

••

Application of biomarker in drug discovery and development.

30

Bakhtiar R. Biomarkers in drug discovery and development. J. Pharmacol. Toxicol. Methods 57(2), 85–91 (2008).

••

Application of biomarker in drug discovery and development.

31

Trusheim MR, Burgess B, Hu SX et al. Quantifying factors for the success of stratified medicine. Nat. Rev. Drug Discov. 10(11), 817–833 (2011).

••

Biomarker program to be an economically valuable model for pharmaceutical development.

13

14

15

Ouellet D, Grossmann KF, Limentani G et al. Effects of particle size, food, and capsule shell composition on the oral bioavailability of dabrafenib, a BRAF inhibitor, in patients with BRAF mutation-positive tumors. J. Pharm. Sci. 102(9), 3100–3109 (2013). Gilmartin AG, Bleam MR, Groy A et al. GSK1120212 (JTP-74057) is an inhibitor of MEK activity and activation with favorable pharmacokinetic properties for sustained in vivo pathway inhibition. Clin. Cancer Res. 17(5), 989–1000 (2011).

16

Verma S, Miles D, Gianni L et al. Trastuzumab emtansine for HER2-positive advanced breast cancer. N. Engl. J. Med. 367(19), 1783–1791 (2012).

32

Parker JL, Lushina N, Bal PS, Petrella T, Dent R, Lopes G. Impact of biomarkers on clinical trial risk in breast cancer. Breast Cancer Res. Treat. 136(1), 179–185 (2012).

17

Bollag G, Tsai J, Zhang J et al. Vemurafenib: the first drug approved for BRAF-mutant cancer. Nat. Rev. Drug Discov. 11(11), 873–886 (2012).

33

Hayashi K, Masuda S, Kimura H. Impact of biomarker usage on oncology drug development. J. Clin. Pharm. Ther. 38(1), 62–67 (2013).

18

Dupart J, Zhang W, Trent JC. Gastrointestinal stromal tumor and its targeted therapeutics. Chin. J. Cancer 30(5), 303–314 (2011).

34

19

Sahu A, Sahu A, Prabhash K et al. Crizotinib: aA comprehensive review. South Asian J. Cancer 2(2), 91–97 (2013).

Desiere F, Romano Spica V. Personalised medicine in 2012: editorial to the special issue of New Biotechnology on “molecular diagnostics & personalised medicine”. N. Biotechnol.29(6), 611–612 (2013).

35

Obermeier M, Symons J, Wensing AM. HIV population genotypic tropism testing and its clinical significance. Curr. Opin. HIV AIDS 7(5), 470–477 (2012).

Chapman PB, Hauschild A, Robert C et al. Improved survival with vemurafenib in melanoma with BRAF V600E mutation. N. Engl. J. Med. 364(26), 2507–2516 (2011).

36

Hoggatt J. Personalized medicine – trends in molecular diagnostics: exponential growth expected in the next ten years. Mol. Diagn. Ther. 15(1), 53–55 (2011).

20

392

Pidugu & Novac




••

Prospective collaboration between pharmaceutical and diagnostic companies.

40

NCBI Database.. www.ncbi.nlm.nih.gov/gene/

37

Bradley E. Incorporating biomarkers into clinical trial designs: points to consider. Nat. Biotechnol. 30(7), 596–599 (2012).

41

ClinicalTrials.gov: Definition of terminology. http://clinicaltrials.gov/ct2/about-studies/glossarytitle

42

38

Marrer E, Dieterle F. Promises of biomarkers in drug development--a reality check. Chem. Biol. Drug Des. 69(6), 381–394 (2007).

ClinicalTrials.gov. http://clinicaltrials.gov/ct2/html/images/info/public.xsd

43

MDXMonitor. www.mdxmonitor.com

39

Research Article

ClinicalTrials.gov. www.ClinicalTrials.gov.



393