Translation of New Molecular Imaging Approaches to the Clinical Setting: Bridging the Gap to Implementation

SEARCH STRATEGY

To gain insight into novel PET tracers being tested in breast cancer, PubMed/Medline was searched, with special attention to studies involving ¹⁸F-FDG, ¹⁸F-FLT, and ¹⁸F-FES tracers. In April 2015, ClinicalTrials.gov was searched for ongoing clinical trials with the search terms [PET] AND [breast cancer]. In total, 164 ongoing PET studies were found.

TRANSITION OF TRACERS FROM PRECLINICAL EVALUATION TO FIRST-IN-HUMAN STUDIES

Bringing tracers from the research-and-development phase to the clinical setting can be a major challenge. Several barriers can cause inefficient translation of novel chemical entities to clinical-grade radiopharmaceuticals; these include lack of good-manufacturing-practice facilities, lack of resources, and insufficient knowledge of the translational process and regulatory requirements.

An investigational radiopharmaceutical for use in a clinical trial is an investigational medicinal product for which an investigational medicinal product dossier (IMPD) is required in Europe. In the United States, the procedure is similar; an investigational new drug application is submitted instead of an IMPD. The application includes data on product quality and safety. The IMPD is submitted together with the clinical trial application to the competent authority. The IMPD outlines the quality and safety of the investigational radiopharmaceutical based on data gathered during the development process.

In the first phase, on the basis of a good rationale, the radiochemical synthesis—including purification, characterization, initial formulation, and stability—is developed. The result of this phase is a development report, which describes the critical process steps and forms the basis for the subsequent technology transfer step. If tracer development is successful and preclinical data are not yet available in the literature, the tracer is evaluated with in vitro and in vivo models to assess its biodistribution and estimate radiation dosimetry. Information from this preclinical evaluation phase is incorporated into the nonclinical pharmacology, pharmacokinetics, and toxicology section and the risk/benefit section of the IMPD. After a decision is made to translate the tracer to the clinical setting, the pharmaceutical or chemistry, manufacturing, and control phase starts. Techniques are transferred from the research-and-development laboratory setting to the good-manufacturing-practice environment. The manufacturing process is described, starting materials are defined, master batch records and testing procedures are documented, and final release specifications and in-process controls are determined and their justification is described. Next, the analytic methods and the manufacturing process are validated, and the subsequent stability of the final drug product is assessed. The results of the pharmaceutical phase are approved master batch and testing records and validation and stability reports. This information is included in the chemical and pharmaceutical section of the IMPD.

If necessary, a toxicology study is performed, and the results are described in the nonclinical pharmacology, pharmacokinetics, and toxicology section. In the last step, all data are reviewed, and the final IMPD is authorized and submitted to the competent authority. A yearly product quality review and an update of the IMPD are mandatory.

Apart from product and process requirements, other essential elements that ensure final product quality are premises and equipment (qualified and monitored clean rooms, laminar flow hoods, and isolator hot cells); well-trained and qualified personnel; and a good-manufacturing-practice-quality system, including documentation and proper deviation and change in management.

The next step is a first-in-human trial, a small pilot study, for proof of concept and safety. When a tracer is proven safe and considered to be of clinical utility, larger studies with quality controls and standardization steps are required to gain the LoE needed to implement the tracer into the clinical setting. Most of the knowledge about the application of tracers in breast cancer trials concerns ¹⁸F-FDG, ¹⁸F-FLT, and ¹⁸F-FES.

ROLE OF ¹⁸F-FDG PET IN STANDARD BREAST CANCER CARE

¹⁸F-FLT PET IN BREAST CANCER

Imaging of cellular proliferation with ¹⁸F-FLT once held great promise for tumor imaging and quantifying a treatment response. However, the facts that the signal intensity is not always high enough and that false-negative and false-positive findings occur result in low sensitivity and specificity in breast cancer (35). On the basis of 11 studies with 189 patients, ¹⁸F-FLT PET is not a strong tool for staging or diagnosing breast cancer because of false-negative results for small axillary lymph nodes. It may play a role in predicting a therapy response. However, the results are equivocal (Table 2). Moreover, it is difficult to pool individual patient data given the different outcome measurements and different imaging methods, labeling procedures, and scan protocols used.

¹⁸F-FES PET IN BREAST CANCER

Therapy selection for breast cancer patients is mainly based on the presence of the ER, the progesterone receptor, and human epidermal growth factor receptor 2 (HER2). Immunohistochemical (IHC) tumor staining for these receptors is considered to be the gold standard (4,30). In the MBC setting, repeated biopsies are advised because receptor expression can change over time. However, a biopsy does not necessarily capture inter- and intratumoral heterogeneity (36,37).

More than 70% of breast cancers overexpress the ER. This fact explains the major interest in the 16 ¹⁸F-FES PET studies performed in over 750 breast cancer patients (Table 3). Six studies investigated the correlation between ER immunohistochemistry and ¹⁸F-FES uptake; the correlation in all of them was good. Predicting the response to endocrine therapy was examined in 8 trials comprising 240 patients. Absence of ¹⁸F-FES uptake predicted the failure of endocrine therapy (38,39), and a decrease in uptake during therapy indicated a response to the antihormonal drugs tamoxifen and fulvestrant (40–43).

The results of 11 new studies including an additional 852 patients are expected over the next few years (Table 4). Pooling individual patient data may provide more solid evidence for the role of ¹⁸F-FES PET in the clinical setting. However, pooling of data may be challenging given the various tracer dosages, time frames, and reconstructions used.

ER-positive and progesterone receptor–positive tumors showed less uptake of ¹⁸F-FDG than hormone receptor–negative tumors (32). The role of ¹⁸F-FES PET imaging in staging has not yet been proven, but with the knowledge that ¹⁸F-FDG PET often shows lower uptake in hormone-positive tumors, it can be hypothesized that ¹⁸F-FES PET may be of help in staging for patients with such tumors. A trial comparing immunohistochemically determined hormonal status with ¹⁸F-FES uptake in MBC patients with hormone receptor–positive or –negative disease before treatment is ongoing (NCT01957332) (44).

MULTICENTER STUDIES AND REPRODUCIBILITY OF RESULTS

When multicenter studies are started, for all steps in the manufacturing process that are conducted at more than one center, evidence that the final drug products and manufacturing processes are comparable should be provided. This goal could be achieved by cross-validation of the manufacturing processes, including quality control. The National Cancer Institute Cancer Imaging Program has been creating investigational new drugs for use as imaging agents. A subset of the documents filed is being made available to the research community to implement the routine synthesis of tracers at various facilities and to assist investigators with the filing of their own investigational new drugs (45).

A prerequisite for a relevant scan or biomarker for the clinical setting is a high degree of test–retest accordance; in addition, the results of the test should be independent of the hospital at which the test is performed. European Association of Nuclear Medicine procedure guidelines have set rules for harmonizing data and obtaining better reproducibility. The American College of Radiology and the European Association of Nuclear Medicine Research Ltd. accreditation programs, the Society of Nuclear Medicine and Molecular Imaging Clinical Trials Network, and the Quantitative Imaging Biomarkers Alliance of the Radiologic Society of North America are all initiatives to make (molecular) imaging a standardized diagnostic modality in clinical medicine and research. Fortunately, interest in and intention to combine European and U.S. guidelines for molecular imaging to gain more uniform data are growing (46). A retrospective assessment of the compliance of 11 sites with an imaging guideline for ¹⁸F-FDG PET, however, showed poor compliance, possibly affecting tumor uptake quantification (47). These data show the need for prospective quality control during studies.

A protocol to guide the upfront performance of ¹⁸F-FDG PET/CT studies within the context of single- and multiple-center clinical trials has been published (48). It provides standards for all phases of imaging in oncological trials. This Uniform Protocol for Imaging in Clinical Trials is another step toward the larger patient datasets and uniform databases that allow individual patient data meta-analysis. In analogy to the database formed for RECIST, data from different trials can be combined, providing the large patient groups required for solid evidence. Although current guidelines and accreditation programs focus on ¹⁸F-FDG PET, similar approaches can be used for the new tracers.

TRIAL DESIGNS TO PROVE ROLES OF NEW MOLECULAR IMAGING METHODS IN CLINICAL SETTINGS

Implementing PET imaging as part of standard care requires proven safety and added benefit beyond existing care. Benefits can include improved patient outcomes as well as reduced costs or physical or emotional burdens on patients. Cost savings could be realized by avoiding surgeries and reducing exposure to ineffective treatments (49).

Implementing PET scanning as a biomarker requires the procedure to score well on criteria such as the REMARK criteria (REporting recommendations for tumor MARKer prognostic studies) (50). These criteria were drafted to guide researchers in reporting their studies for tumor markers in oncology, after it was acknowledged that only a few markers had been adopted into clinical practice. Randomized trials are advised to provide the best LoE in support of a screening or predictive biomarker (50) or to show the actual improved patient outcome of a new diagnostic or prognostic strategy incorporating PET imaging relative to routine care. Unfortunately, standard randomized trials are rarely achievable in the field of predictive markers and molecular imaging because of financial boundaries and the limited capacity of tracer production facilities.

Given these constraints, various approaches to prove the clinical value of PET tracers have been undertaken and can be postulated. In the United States, the National Oncologic PET Registry provided prospective data on the clinical impact in daily practice of over 250,000 ¹⁸F-FDG PET scans (51). It has paved the path to defining relevant indications and reimbursement for ¹⁸F-FDG PET. An international registry prospectively collecting data might be able to prove the role of ¹⁸F-FES PET, with a likely added benefit in cases of clinical dilemmas (52,53). Such observational data, when gathered with rigorous methodology, can provide solid evidence for diagnostic and prognostic purposes. With regard to PET tracers for therapy response prediction, observational data can also provide important initial evidence. MBC patients may be the prime target population for a study of therapy response prediction because of the importance of the timely identification of noneffective treatment leading to progressive disease. In addition, tracer uptake can be linked to response measurements at a metastasis level instead of at a per-patient level to lead to increased statistical efficiency and to allow smaller proof-of-concept studies.

Ideally, after the standardization of procedures, smaller prospective studies with meaningful direct clinical endpoints can be pooled in a database for individual patient data meta-analysis and further validation, enabling the data for each patient to contribute to an increasing evidence base for PET imaging applications. Next, when evidence is deemed sufficient for a new tracer to be implemented as part of standard care, a stepped wedge cluster randomized trial could provide final evidence of benefit while actually taking advantage of the logistical challenges of implementing novel PET technology (54). In such a trial, hospitals are randomized over a certain period of time to the start of implementation, and at the end of this period, all hospitals will have implemented the PET technology. Patient outcome and cost-effectiveness data for the old strategy can then be compared with those for the new strategy in a randomized fashion.

There may not be a clear “one-size-fits-all” approach to evaluating the benefit of molecular imaging (55). A prospective, multicenter observational cohort study is taking place in The Netherlands. Its aim is to evaluate the clinical utility of ¹⁸F-FES, ⁸⁹Zr-trastuzumab, and baseline and early ¹⁸F-FDG PET scans in 200 MBC patients. Endpoints include the correlation between PET scans and (progression-free) survival, cost-effectiveness, and quality of life. Apart from PET scans, other biomarkers, such as circulating tumor cells and DNA as well as tumor DNA and tumor biopsies, are being analyzed (44). These strategies will allow study of the roles of ¹⁸F-FDG, ¹⁸F-FES, and ⁸⁹Zr-trastuzumab PET in relation to those of other potential novel biomarkers and will provide information beyond that provided by the standard of care. Another trial will evaluate the clinical utility of ¹⁸F-FES PET in 99 hormone-positive MBC patients and its possible role, relative to that of ¹⁸F-FDG PET, in predicting a response to therapy (NCT02398773).

Cost-effectiveness can be assessed in comparison with standard options and costs per life-year saved. Data on the cost-effectiveness of ¹⁸F-FDG PET in breast cancer patients are limited. Computer models can be used to conduct cost-effectiveness studies (56). In silico simulation studies could help optimize future studies. Such studies assist with the use of the smallest number of patients and thus with generating the lowest costs to obtain a meaningful response prediction signature. Simulated data not only can provide more information concerning the number of patients needed but also can help define thresholds for outcomes as well as define the optimal statistical analysis approach. The first attempt to assess added benefit in terms of the cost-effectiveness of ¹⁸F-FES PET was made by simulating the follow-up for 5 y of women with ER-positive MBC (57). The total costs for the ¹⁸F-FES PET/CT strategy were higher than those for the standard workup or ¹⁸F-FDG PET/CT. Nonetheless, the total number of performed diagnostic tests was smaller for each of the PET/CT strategies than for the standard workup.

DISCLOSURE

This research was supported by Dutch Cancer Society grant RUG 2012-5565; advanced grant OnQView 676339; and research grants from Roche/Genentech, Amgen, Novartis, Pieris, Servier, and AstraZeneca to the University Medical Center Groningen, University of Groningen. Elizabeth G.E de Vries is co-chair of the RECIST committee and is on the data monitoring committee for Biomarin and the advisory board for Synthon. Geke A.P. Hospers is on the advisory boards for Roche, BMS, MSD, and Amgen. No other potential conflict of interest relevant to this article was reported.