Science & Society

Trends in Molecular Medicine July 2015, Vol. 21, No. 7

References 1 Howe, J.J. (2006) The rise of crowdsourcing. Wired Mag. 14, 1–4 2 Sobel, D. (2007) Longitude: The True Story of a Lone Genius Who Solved the Greatest Scientific Problem of His Time, Walker & Company 3 Filippakopoulos, P. et al. (2010) Selective inhibition of BET bromodomains. Nature 468, 1067–1073 4 Jeppesen, L.B. and Lakhani, K.R. (2010) Marginality and problemsolving effectiveness in broadcast search. Organ. Sci. 21, 1016–1033 5 Brabham, D.C. et al. (2014) Crowdsourcing applications for public health. Am. J. Prev. Med. 46, 179–187 6 Ranard, B.L. et al. (2014) Crowdsourcing: harnessing the masses to advance health and medicine, a systematic review. J. Gen. Intern. Med. 29, 187–203 7 Merchant, R.M. et al. (2014) Hidden in plain sight: a crowdsourced public art contest to make automated external defibrillators more visible. Am. J. Public Health 104, 2306–2312 8 Good, B.M. et al. (2014) The cure: design and evaluation of a crowdsourcing game for gene selection for breast cancer survival prediction. JMIR Serious Games 2, e7

[(1042)TD.ENIM]

9 Ku¨ffner, R. et al. (2015) Crowdsourced analysis of clinical trial data to predict amyotrophic lateral sclerosis progression. Nat. Biotechnol. 33, 51–57 10 Holst, D. et al. (2015) Crowd-sourced assessment of technical skills (CSATS): differentiating animate surgical skill through the wisdom of crowds. J. Endourol. Published online April 13, 2015. http://dx.doi.org/ 10.1089/end.2015.0104 11 King, A.J. et al. (2013) Skin self-examinations and visual identification of atypical nevi: comparing individual and crowdsourcing approaches. Cancer Epidemiol. 37, 979–984 12 Good, B.M. and Su, A.I. (2013) Crowdsourcing for bioinformatics. Bioinformatics 29, 1925–1933 13 Van Mierlo, T. (2014) The 1% rule in four digital health social networks: an observational study. J. Med. Internet Res. 16, e33 14 Gottlieb, A. et al. (2015) Ranking adverse drug reactions with crowdsourcing. J. Med. Internet Res. 17, e80 15 Byrnes, J.E.K. et al. (2014) To crowdfund research, scientists must build an audience for their work. PLoS ONE 9, e110329

[(104)TD.SARIEM]

Clinical application of companion diagnostics Jan Trøst Jørgensen Dx-Rx Institute, 3480 Fredensborg, Denmark

Optimally, any prescription should rely on an in-depth understanding of the disease biology and the mechanism of action of the drug. However, despite the hype about precision medicine in recent years, the prescription of most drugs is still, except for a few anticancer drugs, largely based on ‘trial and error’ and not on solid pharmacogenomic biomarker data. Nearly 15 years ago Spear and colleagues published an article in Trends in Molecular Medicine entitled ‘Clinical application of pharmacogenetics’ [1] which, according to Google Scholar, has been cited more than 350 times. One of the reasons for this popularity is that the article draws our attention to the inherent patient variability to pharmacotherapy, which in most cases seems to be of a significant magnitude. The authors analyzed the efficacy of major drugs within several important disease areas, and their conclusion was that the efficacy rate ranged from 25% to 80%. At the lower end, oncology had an efficacy rate of 25% for cancer chemotherapy, but for most other disease areas the efficacy rate was no higher than 40–60%. The efficacy rates of pharmacotherapy included disease areas such as asthma, rheumatoid arthritis (RA), depression, osteoporosis, diabetes, incontinence, and more. To increase the efficacy and predictability of pharmacotherapy, the authors argued for an implementation of pharmacogenetics testing:

Corresponding author: Jørgensen, J.T. ([email protected]). Keywords: drug efficacy; companion diagnostics; pharmacogenetics; pharmacotherapy; precision medicine; oncology. 1471-4914/ ß 2015 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.molmed.2015.05.003

‘The purpose of a clinical pharmacogenetic assay is to distinguish between those patients who are more and those who are less likely to respond to a drug, or conversely, those who are more and those who are less at risk for adverse events. With this information, better choices for drug therapies can be made to maximize the likelihood of efficacious treatment and minimize the risk for adverse reactions’ [1]. The concept described here is similar to how we understand the use of a companion diagnostic (CDx) assay today, which has gained foothold within oncology in recent years. However, for other diseases drug prescription still seems to be based mainly on the ‘trial and error’ approach. This short article will briefly describe the current concept of CDx and, furthermore, will make a statement on how far we have come with the implementation of the ideas described by Spear and colleagues. Companion diagnostics (CDx) A CDx assay, according to the US FDA guidance document issued in August 2014, is defined as an in vitro diagnostic device that provides information that is essential for the safe and effective use of a corresponding therapeutic product [2]. Furthermore, the FDA specifies several areas in which such an assay could be essential, and which overall can be summarized as outcome prediction (efficacy and safety) as well as therapy monitoring. No doubt, the efficacy predictions of CDx assays have attracted the most attention so far, especially for several targeted anticancer drugs. Here the concept of having a predictive assay in conjunction with a drug was first introduced by the development of trastuzumab (Herceptin) for the treatment of advanced breast cancer [3]. Trastuzumab obtained FDA approval in September 1998 and, on the same day, the immunohistochemical assay 405

Science & Society

Trends in Molecular Medicine July 2015, Vol. 21, No. 7

for HER2 overexpression (HercepTest) was approved. This co-development and co-approval process makes sense because the assay needs to be available at the same time as the drug to guide the prescription process. A CDx assay is linked to a specific drug and is most often developed in parallel to the drug, using the drug/diagnostic co-development model [4]. The success of this model depends very much on the strength of the biomarker hypothesis deduced during the early research and preclinical phases of the development of the drug. The generation of a solid hypothesis requires a thorough molecular understanding of both the disease biology and the drug mechanism of action. During the subsequent development phases, the CDx assay undergoes both analytical and clinical validation. For the final clinical validation, the CDx assay must demonstrate its ability to predict the treatment outcome in the individual patient. A CDx assay will only be deemed useful if it provides information that can discriminate between responding and non-responding patients. Clinical application of companion diagnostics and drug efficacy Optimally, any prescription should rely on an in-depth understanding of the disease biology and the mechanism of the action of the drug. This has yet to be realized, despite pharmacogenetics having been on the agenda for the past 50 years. However, within one disease area, advances in molecular medicine and molecular diagnostics seem to have made an entry into clinical practice – and this is

within oncology. Here, the new molecular understanding has given sufficient insight to allow us to practice a more rational pharmacotherapy, and this has simultaneously led to the development of several CDx assays. When Spear and colleagues made their assessment of the efficacy of cancer chemotherapy in 2001, which was the mainstay for most oncological treatments at that time, their estimate for the efficacy rate was 25%. A similar assessment today of the efficacy rate of the anticancer drugs that subsequently obtained regulatory approval indicates obvious improvements. Table 1 shows the results of a survey performed in relation to some of the anticancer drugs that have obtained approval over the past 15 years for different advanced and/or metastatic cancer diseases. The objective response rates listed in the table are extracted from the prescribing information available to healthcare professionals, and have been retrieved from the US FDA database (http://www.accessdata.fda.gov/ scripts/cder/drugsatfda/), which contains information on all approved prescription and over-the-counter drugs. The majority of the drugs listed in Table 1 are not classified as chemotherapeutics but instead as targeted drugs, either monoclonal antibodies or small molecules – predominantly tyrosine kinase inhibitors. If these drugs are broken down according to whether they have a CDx assay linked to their use or not, there is a clear trend towards a somewhat higher response rate for the former group of drugs. The response rates for this group range from 80.2% to 34.0%, while for the group of drugs that have no CDx assay linked to their use the range is from 6.8% to 45.0%.

Table 1. Objective response rates for anticancer drugs with and without a CDx assay linked to their usea,b Drug Pertuzumab (Perjeta) Crizotinib (Xalkori) Erlotinib (Tarceva) Cetuximab (Erbitux) Ceritinib (Zykadia) Imatinib Mesylate (Gleevec) Dabrafenib (Tafinlar) Afatinib (Gilotrif) Vemurafenib (Zelboraf) Ado-trastuzumab emtansine (Kadcyla) Olaparib (Lynparza) Bevacizumab (Avastin) Ixabepilone (Ixempra) Paclitaxel protein-bound particles (Abraxane) Pemetrexed (Alimta) Pembrolizumab (Keytruda) Ziv-aflibercept (Zaltrap) Cabazitaxel (Jevtana) Sorafenib (Nexavar) Eribulin mesylate (Halaven) Ipilimumab (Yervoy) Sunitinib malate (Sutent)

Indication Breast cancer (HER2+) NSCLC (ALK+) NSCLC (EGFR+) Colorectal cancer (EGFR+/KRAS) NSCLC (ALK+) GIST (CD117+) Melanoma (BRAF+) NSCLC (EGFR+) Melanoma (BRAF+) Breast cancer (HER2+)

CDx Assay(s) HercepTest (Dako)/HER2 IQFISH pharmDx (Dako) Vysis ALK Break Apart FISH probe kit (Abbott) Cobas EGFR mutation test (Roche) EGFR pharmDx (Dako)/KRAS RGQ PCR kit (Qiagen)

Platform IHC/FISH FISH PCR IHC/PCR

Response rate 80.2% 65.0% 65.0% 57.0%

Vysis ALK Break Apart FISH probe kit (Abbott) c-Kit pharmDx (Dako) ThxID BRAF kit (BioMe´rieux) EGFR RGQ PCR kit (Qiagen) Cobas 4800 BRAF V600 mutation test (Roche) HercepTest (Dako)/HER2 IQFISH pharmDx (Dako)

FISH IHC PCR PCR PCR IHC/FISH

54.6% 53.9% 52.0% 50.4% 48.4% 43.6%

Ovarian cancer (BRCA+) Colorectal cancer Breast cancer NSCLC

BRACAnalysis CDx (Myriad) No CDx No CDx No CDx

PCR

34.0% 45.0% 34.7% 33.0%

NSCLC Melanoma Colorectal cancer Prostate cancer Thyroid carcinoma Breast cancer Melanoma GIST

No No No No No No No No

CDx CDx CDx CDx CDx CDx CDx CDx

27.1% 24.0% 19.8% 14.4% 12.0% 11.0% 10.9% 6.8%

a

The indications mentioned in the table are for metastatic and/or advanced-stage disease. All the drugs listed in the table obtained FDA approval after 2001 (http://www. accessdata.fda.gov/scripts/cder/drugsatfda/).

b Abbreviations: FISH, fluorescence in situ hybridization; GIST, gastrointestinal stromal tumor; IHC, immunohistochemistry; NSCLC, non-small-cell lung carcinoma; PCR, polymerase chain reaction.

406

Science & Society Making a comparison as outlined in Table 1 may not be fair because differences in the indications, line of therapy, and patient populations may influence the results. However, despite these reservations there seems to be a clear trend in favor of drugs that have a CDx assay linked to their use. Hence, for cancer diseases the conclusion must be that pharmacogenetic testing matters for the response to pharmacotherapy. Despite this apparent success, there is no reason to be overenthusiastic in relation to oncology either. According to CenterWatch, 112 oncology drugs have obtained FDA approval since 2001, and only 14 of these drugs have a CDx assay linked to their use (CenterWatch, FDA Approved Drugs for Oncology. https://www. centerwatch.com/drug-information/fda-approved-drugs/ therapeutic-area/12/oncology; FDA, List of Cleared or Approved Companion Diagnostic Devices. http://www.fda. gov/MedicalDevices/ProductsandMedicalProcedures/ InVitroDiagnostics/ucm301431.htm). What is the status of the other disease areas that Spear and colleagues [1] described in their article? Over the past 15 years, despite several drug approvals within different disease areas such as type II diabetes, depression, RA, and more, no CDx assay has been co-developed for any of them. The introduction of the anti-TNF (tumor necrosis factor) drugs, such as infliximab (Remicade) and adalimumab (Humira), has significantly improved the outcome for RA patients; however, the use of these drugs is essentially based on a trial-and-error approach and not all patients respond adequately. CDx assays for this relatively expensive group of drugs would be of great value, and hopefully such assays will be developed in the near future [5]. In addition to the few anticancer drugs for which the US FDA have approved a CDx assay, there is a somewhat longerlist containing close to 120 additional drugs that have pharmacogenomic biomarker information included in the drug labeling (FDA, Table of Pharmacogenomic Biomarkers in Drug Labeling. http:// www.fda.gov/%20drugs/scienceresearch/researchareas/ pharmacogenetics/ucm083378.htm). This information could be important in relation to both identifying drug response and avoiding side effects in the individual patient. However, the assays needed for these pharmacogenomic

[(104)TD.ENIM]

Trends in Molecular Medicine July 2015, Vol. 21, No. 7

biomarkers are seldom clinically accessible to the same extent as the CDx assays, and their application in standard care is still a long way off [6]. Concluding remarks Despite the hype about precision medicine in recent years, the prescription of most drugs is still largely based on trial and error and not on solid pharmacogenomic biomarker data. Such an approach can have serious medical consequences for the individual patient as well as economic consequences for the healthcare system. For most serious chronic diseases, early diagnosis and early intervention are two elements of key importance, and the intervention needs to be correct. So far, CDx assays have been reserved for oncology, and have improved the efficacy for several drugs. No doubt, more widespread use of pharmacogenomic biomarkers would lead to a more rational and cost-effective pharmacotherapy to the benefit of both the individual patient and the healthcare system as a whole. However, apart from anticancer drugs, the pharma and biotech companies do not appear to have prioritized pharmacogenomic biomarkers, and for a more general use of CDx assays there is still a long way to go. Disclaimer statement Jan Trøst Jørgensen is working as a consultant for Dako/Agilent and Euro Diagnostica, and has given lectures at meetings sponsored by AstraZeneca, Merck Sharp & Dohme, and Roche.

References 1 Spear, B.B. et al. (2001) Clinical application of pharmacogenetics. Trends Mol. Med. 7, 201–204 2 US FDA (2014) Guidance for Industry and Food and Drug Administration Staff. In Vitro Companion Diagnostic Devices, US Department of Health and Human Services 3 Slamon, D.J. et al. (2001) Use of chemotherapy plus a monoclonal antibody against HER2 for metastatic breast cancer that overexpresses HER2. N. Engl. J. Med. 344, 783–792 4 Olsen, D. and Jørgensen, J.T. (2014) Companion diagnostics for targeted cancer drugs – clinical and regulatory aspects. Front. Oncol. 4, 105 5 Gibson, D.S. et al. (2015) Current and future trends in biomarker discovery and development of companion diagnostics for arthritis. Expert Rev. Mol. Diagn. 15, 219–234 6 Carr, D.F. et al. (2014) Pharmacogenomics: current state-of-the-art. Genes 5, 430–443

407

Clinical application of companion diagnostics.

Optimally, any prescription should rely on an in-depth understanding of the disease biology and the mechanism of action of the drug. However, despite ...
120KB Sizes 0 Downloads 19 Views