Understanding of the term ‘Precision Medicine’ is variable. For
clinicians, pharmacists and clinical pharmacologists, it refers to the
right decision (treat or not), right drug, right combination, right
timing and right dose, taking into account the clinical trial data for
the patient group being treated, and finessing that to the individual
biological and pharmacological variables either evident, or likely to be
evident based on knowledge about comorbidity and body size.
To scientists, precision medicine can mean either developing a new
molecule to fit a specific target or cell of interest, or ‘finding’
based on mathematical and chemical profiling existing drugs that might
‘fit’ the target of interest (1). However, this focus on ‘targets’ as
the method to improve precision has seemingly ignored the well-known
principles of radiobiology, cancer biology and pharmacology, and has not
delivered the improved health outcomes expected (2). Similarly, since
2015, U.S. and other government’s multi-billion-dollar investments into
‘precision’ drugs “delivering the right treatments, at the right time,
every time to the right person whilst helpful to understand some aspects
of disease pathophysiology, has also fueled this single target focus
(3).
In this Themed issue discussion on aspects of science and medicine
people refer to as precision medicine approach are covered. Although
varied, much discussion is related to the importance of understanding
the way each individual both handles and responds to a drug, and
specific dose. How to implement individualized dosing into practice
however can be an even more challenging area, with different levels of
precision costing different amounts, with a spectrum of clinical and
economic benefits.
Added to this difficulty is the systematizing of regimen and dose for
specific cancers which have come into oncology practice over the last
few years. Moynihan et al argue that the financial dependence of
research on industry funding creates a “sponsorship bias” that
overplays efficacy and underplays toxicity. This was confirmed in a
systematic review comparing industry sponsored with independently funded
trials (4).
This is not just an issue with interpretation of results but goes back
to trial design. The choice of a comparator may make the outcome of the
study drug more favorable (4). Over the years in cancer trials, there
has certainly been a move from the principle of treating until maximum
response and then allowing a patient time without symptoms or side
effects of treatment to studies which are designed to continue treatment
until relapse or unacceptable toxicity. This maximizes drug use but is
there evidence that prolonging use of a drug maximizes outcomes?
Maintenance strategies in metastatic disease can be simply continuing
the drug used in induction or switching to another drug, which really
can be considered as early second line treatment (5). Although
established in lymphomas, randomized studies of continuous maintenance
therapy in non-small cell lung cancer (NSCLC) led to no improvement in
response or survival and similarly for survival in colorectal cancer
(5,6). Prolonging first line therapy in breast cancer only showed a
marginal survival benefit. There are ongoing studies, but prolonging
induction therapy lacks the evidence for widespread adoption. Switching
to another drug or targeted therapy, as an early second line treatment,
has shown prolonged survival over induction alone in several tumour
types (5).
The duration of targeted therapies provides examples of how trial design
influences practice. The initial studies of trastuzumab in adjuvant
breast cancer initially reported at the American Society of Clinical
Oncology in 2005 were designed to give 12 months of therapy and that
became the standard of care (7). An ongoing study at the time was
comparing 12 months with 24 months. (8). However, an independent French
group looked at 6 months compared to 12 months and it is only in 2019
that the final analysis could not show non-inferiority of 6 months of
treatment (9). Meanwhile the FinHER study in Finland showed the efficacy
and cost-effectiveness of only 9 weeks of adjuvant trastuzumab after
chemotherapy (10, 11).
In terms of the dose given, the current common method of dosing
cytotoxic drugs is based on a patient’s body surface area (BSA). This
can be inaccurate with considerable variation between patients because
patient-related factors such as organ function, age, gender, activity of
metabolizing enzymes, drug resistance and concomitant drugs can
influence the pharmacokinetics and pharmacodynamics (12). This gives
rise to pharmacologically-based dosing being explored to make individual
patient dosing more precise.
The first step, however, in the treatment of cancer is to select the
drug or drugs most likely to be effective. In the era of precision
medicine what is being investigated is identifying mutations in genes or
changes in the expression of genes or proteins specific to a tumour,
which can be targeted by therapeutics (13). Detecting multiple genomic
changes has been made possible by technological advances like next
generation sequencing (NGS) replacing older single gene testing. Use
multiple platforms combining sequencing of DNA with RNA sequencing and
with the more established techniques such as immunohistochemistry (IHC)
maximizes the potential to discover druggable targets (14). IHC detects
changes at the protein level that reflect gene amplifications such as
HER-2 in breast cancer, gastric and colorectal cancer which can be
targeted with trastuzumab. Rearrangements include EML4-ALK translocation
in non-small cell lung cancer which can be targeted by drugs such as
crizotinib. Now there is testing for biomarkers related to PD-L1
expression in various tumours which can be targeted by checkpoint
inhibitors such as atezolizumab.
Molecular profiling which allows matching treatments for cancers to
their targets has resulted in a boost for new drug development in rare
cancers by using drugs targeted to molecular biomarkers that they have
in common with more common cancers, in basket trials (15).
Limitations of gene expression profiling and IHC can be illustrated in
diffuse large B-cell lymphoma (DLBCL) as summarized by Ofori et al in
this issue of the Journal (16). Its subtypes defined by the cell of
origin can predict survival and response to chemotherapy, and were
initially classified by gene expression profiling. However, fresh tissue
had to be available and often only major centers had the capability. IHC
methods were subject to observer error. The other issue was that serial
tumour profiling during treatment may be able to detect emerging
resistance but serial biopsies may not be feasible and surveillance post
treatment was only available by imaging, which has not been shown to be
associated with a survival benefit.
Liquid biopsies have the potential for solving these issues. Biomarkers
in blood or other body fluids can be identified, including from
circulating tumour DNA (ctDNA), circulating tumour cells (CTC) or
exosomes. Circulating tumour DNA fragments are shed from tumour cells
and show mutations and methylation profiles of the tumour which could be
used to identify targets or predict recurrence in tumours such as DLBCL
(17).
Circulating tumour cells are derived from primary tumours or their
metastases and either actively or passively enter the circulation. Their
DNA, RNA and proteins could be used to discover the molecular profile of
the tumour and their numbers can correlate with treatment outcome (16).
De Souza et al have identified technological and interpretive challenges
to overcome before CTCs are used routinely in the clinic (18).
Exosomes, formed when a cell membrane buds off with contents including
protein, nucleic acids, sugars and lipids are taken up by other cells
and represent communication between cells. They can circulate in many
body fluids. The potential advantage of exomes is their abundance in the
fluids and their contents that may reveal multiple biomarkers which in a
disease such as DLBCL could be used to subtype and therefore predict
prognosis, be used as surveillance during therapy and reveal resistance
mechanisms. They could then be used to follow-up post treatment (16).
A further tool to guide dosing of anti-cancer drugs and predicting
toxicity and response prior to their administration is pharmacogenomics
reviewed by Carr et al (19). Examples with the strongest evidence are
assaying for dihydropyrimidine dehydrogenase (DPYD) which is the rate
limiting enzyme for 5-FU metabolism, encoded by a gene with multiple
variants. Identifying the variant alleles of thiopurine
methyltransferase (TMPT) can identify a low activity genotype which
metabolizes 6-mercaptopurine to an inactive mercaptopurine resulting in
less metabolism of 6-MP to toxic thioguanine nucleotide metabolites.
There are many other potential applications of pharmacogenomics, but
with equivocal or less evidence. The more widespread use of NGS will
allow easier identification of rarer mutations associated with adverse
drug reactions. The lack of routine use of pharmacogenomics is
multifactorial including the expense, accessibility, the time for
processing and the complex interactions including between genomics,
clinical factors and the microbiome which account for the individual
variations (19).
Personalised drug dosing is important in oncology to prevent overdosing,
which otherwise may only become evident when a patient develops severe
side effects, or underdosing resulting in lack of efficacy, which may
not be revealed until scans show a lack of tumour response . The
evidence for some drugs that drug exposure is related to efficacy and
toxicity allows for therapeutic dose monitoring (TDM) such as is used
for dosing antimicrobials. Some examples of attempting TDM with
cytotoxics illustrate the challenges.
For intravenous 5-FU while DPYD genotyping is useful, more precision is
needed for bolus and infusional regimens in a variety of cancers,
including head and neck and colorectal cancer. In articulating the
importance of TDM dosing Schneider JJ et al. in this themed issue,
reiterate that 5-FU dosing by BSA only results in 20-30% patients
achieving the therapeutic range. Exploring the relationship between 5-FU
area under the curve (AUC) and a target dose resulted in the
recommendation of a therapeutic exposure range of 20-30mgh/L for 46-hour
infusion schedules (20). Unfortunately, data is lacking to apply TDM
dosing to the oral prodrug capecitabine, which is just as effective as
5-FU but better tolerated.
Other common cytotoxic drugs are more problematic. Muth et al reviewed
the taxanes; paclitaxel, docetaxel, nab-paclitaxel and cabazitaxel which
illustrate some of the complexities of TDM dosing (21). Paclitaxel which
is commonly dosed weekly or 3 weekly has non-linear pharmacokinetics,
undergoes hepatic metabolism and biliary excretion and there are
interactions with its solvent cremophor, however the time above a plasma
concentration of 0.05 µmol/L does predict neutropenia and polyneuropathy
and may be associated with a favourable clinical outcome, making TDM
dosing desirable. Docetaxel, is also extensively metabolised in the
liver, has linear kinetics, but is formulated with polysorbate 80 rather
than cremophor. Weekly docetaxel has a more favourable toxicity profile
that 3-weekly dosing but it is AUC that predicts febrile neutropenia,
mucositis and diarrhoea. Less research has been done than with
paclitaxel but a small randomized study of TDM and target concentration
intervention (TCI) compared to BSA didn’t show a clear advantage for TDM
and TCI for docetaxel (22). Unfortunately, is no prospective TDM data
for carbazitaxel or nab-paclitaxel.
Early in the development of carboplatin the relationship between drug
exposure and efficacy and toxicity was established and dosing was more
accurately based on renal function (glomerular filtration rate- GFR)
than BSA. However, for specific groups such as infants, anephric
patients and those receiving high-dose carboplatin, TDM dosing is more
desirable that dosing based on (GFR) as summarized by Barnett S et al
(23).
For TDM to be translated into clinical practice, the evidence base must
expand, and sampling strategies need to be simplified, perhaps by micro
sampling such as using dried blood spots or using body fluids other than
blood. There must be better access to TDM laboratories, and the
provision of clinical decision support for interpreting the results of
pharmacometrics which use Bayesian estimations to combine
pharmacokinetics, individual patient characteristics and drug
concentrations (24).
Finally, a barrier which must be addressed to allow clinical translation
of TDM is the demonstration of its economic efficacy which Vithanachchi
DT et al present in a descriptive review (25). They reviewed 11 studies
and noted that only a few drugs have been studied. However, all studies
reviewed found TDM to be cost effective, based on established
incremental cost-effectiveness ratios. In future newer therapeutics
should have an economic analysis of TDM, incorporating the associated
clinical evidence, which in the short term is reduced toxicity and the
long term, a survival advantage.
REFERENCES1. Martin JH, Bowden NA. Drug Repurposing – Overcoming the
translational hurdles to clinical use. Pharmacol Res Perspect 2019, Nov
26, https://doi.org/10.1002/prp2.548.
2. Fay M, Head R, Martin J. Where is the radiobiology and pharmacology
research to improve outcomes on glioblastoma? J Neurooncol 2015, 124:
1-3.
3. Obama B. The Precision Medicine Inititive
https://obamawhitehouse.archives.gov/precision-medicine [Last
Accessed 11 Dec 2020]
4. Moynihan R, Beros L, Hill S et al. Pathways to independence: towards
producing and using trustworthy evidence. BMJ 2019, Dec 3; 367:16576.
Doi: 10.1136/bmj.e3502.
5. Rowinski E, Fournel P, Bernichon E, Bouleftour W, Magné N, Mery B.
Maintenance therapy in metastatic solid tumours. Innovative strategy or
simply second-line treatment? Am J Clin Oncol 2019, 42: 615-623.
6. von Plessen C, Bergman B, Anderson O et al. Palliative chemotherapy
beyond three courses conveys no survival or consistent quality-of-life
benefits in advanced non-small-cell lung cancer.Br J Cancer 2006, 95:
966-973
7. Perez EA, Romond EH, Suman VJ et al. Four-year follow-up of
trastuzumab plus adjuvant chemotherapy for operable human epidermal
growth factor receptor positive breast cancer: joint analysis of data
from NCCTG N9831 and NSABP B-31. J Clin Oncol 2011, 29: 3366-3372.
8. Cameron D, Piccart-Gebhart MJ, Gelber RD et al. 11 years’ follow-up
of trastuzumab after adjuvant chemotherapy in HER2-postive early breast
cancer: final analysis of the HERceptin Adjuvant (HERA trial. Lancet
2017, 389: 1195-1205.
9. Pivot X, Romieu G, Debled M et al. 6 months versus 12 months of
adjuvant trastuzumab in early breast cancer (PHARE): final analysis of a
multicenter, open-label, phase 3 randomized trial. Lancet 2019, 393:
2591-2598.
10. Joensuu H, Kellokumpu-Lehtinen PL, Bono P et al. Adjuvant docetaxel
or vinorelbine with or without trastuzumab for breast cancer. N Engl JK
Med 2006, 354: 809-820.
11. Purmonen TT, Pänkäläinen E,Turunen JH, Asseburg C, Martikainen JA.
Short-course adjuvant trastuzumab therapy in early stage breast cancer
in Finland: cost-effectiveness and value of information analysis based
on the 5-year follow-up results of the FinHer trial. Ann Oncol 2011, 50:
344-352.
12. Kaestner SA, Sewell GJ. Chemotherapy dosing part I: scientific basis
for current practice and use of body surface are. Clin Oncol (R Coll
Radiol) 2007, 19: 23-37.
13. Malone ER, Oliva M, Sabatini PJB, Stockley TL, Siu LL. Molecular
profiling for precision cancer therapies. Genome Med 2020 12:8
https://doi.org/10.1186/s13073-019-0703-1. [Last accessed Jun 10
2020].
14. Zimmer K, Kochner F, Spizzo G, Salem M, Gastl G, Seeber A. Treatment
according to molecular profiling in relapsed/refractory cancer patients:
a review focusing on latest profiling studies. Comput Struct Biotechnol
J 2019, 17: 447-453.
15. Wang S, Chen R, Tang Y et al. Comprehensive genomic profiling of
rare tumour: routes to targeted therapies. Front Oncol 2020,
https://doi.org/10.3389/fonc.2020.00536 [Last accessed 10 Jun
2020]
16 Ofori K, Bhagat G, Rai A. Exosomes as liquid biopsy biomarkers in
diffuse large B-cell lymphoma (DLBCL) – Current. State-of-the -art and
unmet needs. Br J Clin Pharmacol 2020, X: XX-XX
17. Roschewski M, Dunleavy K, Pittluga S et al. Circulating DNA and CT
monitoring in 124 patients with untreated diffuse large B-cell lymphoma:
a correlative biomarker study. Lancet Oncol 2015, 16: 541-549.
18. De Souza P, Po J, Scott K et al. Can on study changes in circulating
tumour cell (CTC) counts be used as a predictive measure for therapeutic
efficacy in caner clinical trials? A review of the literature. Br J Clin
Pharmacol 2020, X: XX-XX.
19 Carr DF, Turner RM, Pirmohamed M et al. Pharmacogenomics of
anticancer drugs: personalising the choice and dose to manage drug
response. Br J Clin Pharmacol 2020, XX: XX-XX.
20. Schneider J, Galettis P, Martin J. Overcoming barriers to
implementing precision dosing with 5-Fluorouracil and capecitabine. Br J
Clin Pharmacol 2020, X:XX-XX.
21 Muth M, Ojara FW, Joerger M. Role of TDM-based dose adjustments for
major taxane anticancer drugs. Br J Clin Pharmacol 2020, X: XX-XX.
22. Engel FK, Loos WJ, van der Bol JM et al. Therapeutic drug monitoring
for the individualization of docetaxel dosing: a randomised
pharmacokinetic study. Clin Cancer Res 2011, 17: 353-362.
23 Barnett S, Kong J, Makin G, Veal GJ. Over a decade of experience with
carboplatin therapeutic drug monitoring in a childhood cancer setting in
the United Kingdom. Br J Clin Pharmacol 2020, X: XX-XX.
24 Menz BD,, Stocker SL, Verougstraete N et al. Barriers and
opportunities for the clinical implementation of therapeutic drug
monitoring in oncology. Br J Clin Pharmacol 2020, X: XX-XX.
25. Vithanachchi D, Maujean A, Downes MJ, Scuffham P. A systematic
review of economic evaluations of therapeutic drug monitoring
interventions for cancer treatments. Br J Clin Pharm 2020, X: XX-XX.