⁸⁹Zr-ImmunoPET Companion Diagnostics and their Impact in Clinical Drug Development

Abstract

Therapeutic monoclonal antibodies (mAbs) have been used in cancer treatment for 30 years, with around 24 mAb and mAb:drug conjugates approved by the FDA to date. Despite their specificity, efficacy has remained limited, which, in part, derails nascent initiatives towards precision medicine. An image-guided approach to reinforce treatment decisions using immune positron emission tomography (immunoPET) companion diagnostic is warranted. This review provides a general overview of current translational research using Zr-89 immunoPET and opportunities for utilizing and harnessing this tool to its full potential. Patient case studies are cited to illustrate immunoPET probes as tools for profiling molecular signatures. Discussions on its utility in reinforcing clinical decisions as it relates to histopathological tumor assessment and standard diagnostic methods, and its potential as predictive biomarkers are presented. We finally conclude with an overview of practical considerations to its utility in the clinic.

Introduction

Therapeutic monoclonal antibodies (mAbs) gained clinical utility in 1985 with the first US Food and Drug Administration (FDA)-approval of the biologic, muromonab-CD3 (Orthoclone™ OKT3), specific for cluster of differentiation (CD3), a co-receptor present on all T-cells¹. Since then, applications in cancer have been exploited with the approval of rituximab (anti-CD20) in 1997² followed by trastuzumab (anti-HER2) in 1998³. By 2016, there were 24 mAbs and antibody drug conjugates (ADC) approved by the FDA for cancer treatment. These mAbs are directed to a specific target ranging from tumor and cell-surface associated antigens to biomarker signatures within the tumor microenvironment. Despite their specificity and moderate safety profile, clinical efficacy of these mAbs remains limited due to perpetuating factors, including but not limited to i) unpredictable tumor antigen density, ii) internalizing status of the mAb:antigen complex, iii) target hit rate, iv) vascular penetration, and, v) tissue distribution, which may impact adverse events^4–8. All of these factors underscore the need for precision medicine, borne out of the intent of tailoring the disease treatment and prevention by providing the right drug to the right patient at the appropriate time and dose. A logical approach to precision medicine explores non-invasive imaging tools that can be repeatedly utilized to profile tumors at the molecular level, and to augment flaws present in biopsies. With this perspective, antibody or immune-based positron emission tomography (immunoPET) was developed to provide a direct readout of antigen density present within each lesion; moreover, the pharmacokinetic and dosimetric properties of the mAb, in the case of radioimmunotherapy, can be considered cognate when compared to the imaging tool⁹. Taken together, immunoPET has a high potential to influence and direct informed decisions in drug design and development. In this review, we present a general overview of immunoPET, specifically imaging agents labeled with ⁸⁹Zr while briefly touching on tracer development. We discuss the clinical impact of ⁸⁹Zr-immunoPET, its role in drug development and factors to consider to harness its utility as an effective companion diagnostic (CDx).

ImmunoPET tracer development

The development of immunoPET tracers relies on the following principles: i) the biological and chemical properties of the mAb, ii) the radionuclide chosen iii) the chelate selected, and iv) the stability of the linker between mAb and chelate. MAbs for patient use are either humanized or made fully human to prevent human anti-mouse antibody response (HAMA)¹⁰. The size of full-length biologics (~150 kDa) prolongs their half-life in the blood, which affects the length of delivery to the tumor target and clearance from healthy tissues. Thus, pairing with long-lived radionuclides ⁶⁴Cu (t_1/2 ~ 12.7 h), ⁸⁶Y (t_1/2 ~ 14.7 h), ⁸⁹Zr (t_1/2 ~ 78.4 h), and ¹²⁴I (t_1/2 ~ 100.3 h) is the most common strategy¹¹. Matching the physical and biological half-lives of the PET nuclide and the mAb, respectively, ensures that the probe accumulates in the tumor before the radioactivity decays and further allows clearance from normal tissues. In this regard, enhanced signal-to-noise ratio – one of the primary considerations in diagnostic imaging – is achieved.

One limitation to using full mAbs specifically for imaging purposes is the long wait times between tracer administration and imaging acquisition, as well as non-trivial dose exposure of non-target organs. Tracer pharmacokinetics can be improved by decreasing its size, effectively reducing circulation time, and minimizing dose exposure to the patient¹². With this perspective, smaller fragment constructs are engineered offering shorter blood residencies and faster tumor target delivery¹⁰. These fragments mostly retain the variable region where the antigen-binding site is primarily located. Suggested PET radionuclide tags to complement mAb fragments are provided in Table 1. Moderately-sized fragments (i.e. F(ab)’2 (~100–110 kDa), minibody (~75 kDa), and diabody (~50 kDa)) may be appropriately labeled with ¹⁸F (t_1/2 ~ 109 min), ⁶⁴Cu (t_1/2 ~ 12.7 h) and ⁸⁶Y (t_1/2 ~ 14.7 h). Smaller-sized fragments like affibodies (~ 6 kDa), nanobodies or single domain antibodies (~12–15 kDa) can be radiolabeled with shorter-lived isotopes like ¹⁸F and ⁶⁸Ga (t_1/2 ~ 68 min), which consequently decreases the radiation exposure of the patient¹³. The caveat herein lies in the overall rate of clearance and nuclide site delivery of the mAb fragments.

Table 1.

Different antibody fragments and recommended PET radionuclide for companion diagnostic development.

Antibody Fragments	Recommended PET Nuclide	References
Affibody (~7 kDa), Nanobody (~12–15 kDa)	68Ga 18F 64Cu	69–73
Diabody (~55 kDa)	18F 64Cu 89Zr	74–76
Minibody (~80 kDa)	64Cu 89Zr	77–78
Fab’2 (~100–110 kDa)	64Cu 89Zr 124I	79

Zirconium-89 immunoPET tracers

Standardized production and commercial availability has made the development of Zr-89 radiolabeled mAbs relatively straightforward^14–15. As a radiometal, Zr-89 requires complexation to prevent random, non-specific binding to non-targeted tissue (usually the bone), which consequently lowers contrast¹⁶. To date, only desferrioxamine (DFO), a known iron-sequestering siderophore with three hydroxamate groups is currently utilized as a chelate despite reports of metal:complex in vivo instability^17–18. DFO bioconjugation techniques were established either through non-specific attachment to terminal lysines^19–20 and cysteines¹³ or through a more discriminate glycan selective labeling²¹. Consequently, preclinical research flourished with many imaging probes developed to target different oncogenic molecular signatures. A significant number of these tracers were developed to target surface-bound biomarkers, such as i) members of the epidermal growth factor receptor family (e.g. EGFR²², HER2²³ and HER3²⁴), ii) prostate-specific membrane antigen (PSMA)²⁵, iii) prostate stem cell antigen (PSCA)²⁶, iv) CD20²⁷, v) CD44²⁸, vi) programmed death receptor (PD1)²⁹ and vii) programmed death ligand 1 (PD-L1)³⁰, to name a few. Imaging probes targeting secreted signaling proteins (e.g. VEGF, granzyme B, interferon-γ)^31–33, antigen/receptors bound to T cells (e.g. CD3³⁴, CD8³⁵) and shed antigens (e.g. CA19.9³⁶, carcinoembryonic antigen or CEA³⁷) were also investigated. With substantial preclinical data, a number of these tracers have progressed to clinical trials. The first study of a ⁸⁹Zr-mAb probe (⁸⁹Zr-cmAb U36) targeting CD44v6 in patients with head and neck cancer was reported in 2006³⁸. The number of ⁸⁹Zr-based immunoPET probes in the clinic tripled in 2013¹⁷ As of this writing, to the best of our knowledge and after extensive search at clinicaltrials.gov, there are ~22 ⁸⁹Zr-mAbs that are currently undergoing or have completed patient trials. An overview of ⁸⁹Zr-based immunoPET probes can be found in Table 2.

Table 2.

List of ⁸⁹Zr-immunoPET tracers that advanced to clinical trials

ANTIBODY	TARGET	INDICATIONS	CLINICALTRIALS.GOV IDENTIFIER	PHASE AND STATUS
Trastuzumab	HER2	Metastatic HER2+ Breast cancer	NCT01420146	Phase 1, Completed
		Metastatic HER2+ Breast cancer; to select patients for T-DM1 treatment	NCT01565200	Phase 2, Active, not recruiting
		Unsuspected HER2 Breast Metastases	NCT02286843	Recruiting
		Trastuzumab-resistant Breast Cancer; measure HER2 post-treatment with HSP90 inhibitor AUY922	NCT01081600	Phase 1/2, Completed
		Esophagogastric cancer	NCI-2016–00986, NCT02023996	Phase 1; Recruiting
		HER2+ primary malignancy	NCT03109977	On-going but not recruiting
Ibritumomab Tiuxetan	CD20	Non-hodgkins lymphoma		Complete
Bevacizumab	VEGF	Inflammatory Breast Cancer	NCT01894451	Phase 1; recruiting
		Pulmonary arterial hypertension	NCT03166306	Phase 1/2; not open to recruitment at the time of writing
huJ591	PSMA	Prostate cancer	NCT02693860	Phase 1; recruiting
		Metastatic prostate cancer	NCT01543659	Phase 1/2; On-going but not recruiting
		Glioblastoma	NCT02410577	Phase I, On-going but not recruiting
Girentuximab	Carbonic Anydrase IX	Renal cell carcinoma	NCT02883153	Phase 2/3, Completed
Cetuximab	EGFR	Stage IV cancer	NCT00691548	Phase 1; Completed
		Colorectal cancer	NCT01691391	Completed
Ipilimumab	CTLA-4	Melanoma	NCT03313323	Phase 2; Recruiting
Fresolimumab (GC1008)	TGF-β	Primary brain tumor	NCT01472731	Phase 2; Completed
U36	CD44v6	Head and neck cancer		Completed
Pertuzumab	HER2	HER2 positive malignancy	NCT03109977	Phase 1; On-going but not recruiting
IAb2M	PSMA	Metastatic prostate cancer	NCT01923727	Phase 1/2, Completed
		Prostate cancer, pre-prostatectomy	NCT02349022	Phase 2, Completed
IAb22M2C	CD8	Non Small Cell Lung Cancer, Small Cell Lung Cancer, Squamous Cell Carcinoma Head and Neck, Melanoma, Merkel Cell Tumor, Renal, Bladder, Hepatocellular, Triple Negative Breast, or Gastroesophageal Cancer, Hodgkin’s Lymphoma	NCT03107663	Phase 1, Recruiting
Rituximab	CD20	Lung disease, interstitial pneumonitis	NCT02251964	Phase 2/3; On-going but not recruiting
GSK3128349 (Albumin domain binding antibody)	Albumin	Drug related side effects and adverse reactions	NCT02829307	Phase 1; completed
MPDL3280	PD-L1	Breast cancer, bladder cancer and non small cell lung cancer	NCT02453984	Phase 1; recruiting
Pembrolizumab	PD-1	Non small cell lung cancer	NCT03065764	Phase 2; on-going but not recruiting
GSK2849330	HER3	Solid tumors	NCT02345174	Phase 1; Completed
AMG211	HER3	Advanced gastrointestinal cancer	NCT02760199	Phase 1; Completed
RO5479599	HER3	Metastatic and/or Locally Advanced Malignant HER3-Positive Solid Tumors of Epithelial Cell Origin	NCT01482377	Phase 1; Completed
MMOT0530A	Mesothelin	Unresectable pancreatic cancer, platinum-resistant ovarian cancer	NCT01832116	Phase 1; Completed
MSTP2109A	STEAP1	Prostate cancer	NCT01774071	Phase 1/2; On-going but not recruiting
HuMab-5B1 (MVT-2163)	CA19.9	Pancreatic Cancer; tumors that express CA19.9	NCT02687230	Phase 1; Recruiting

The Clinical Impact of ImmunoPET Companion Diagnostics

Molecular profiling of lesions

Understanding the molecular profile of the malignancy is necessary to determine treatment indications. A standard clinical strategy obtains tumor specimens through surgical or core needle biopsies in solid tumors for histopathological analyses. Liquid biopsies are also obtained from blood, urine, sputum, or cerebrospinal fluid, which carry shed or circulating biomarkers³⁹. Biopsy-driven molecular profiling is often fraught with problems and disadvantages since access to the tumor sites may be difficult, often requiring complicated invasive procedures⁴⁰. Tumor heterogeneity renders biopsies inconsistent, which can inadequately portray the presence and level of expression of the molecular signature; thus, requiring more tests to accurately characterize the tumor. Consequently, proper histopathological analysis of the receptor/antigen density may not be reflected, potentially eliminating a patient from benefiting from molecular-based treatments. Repeat biopsies are performed on patients to pathologically confirm malignancy to direct treatment decisions, but secondary biopsy results may not match the original pathology report⁴¹. Moreover, multiple sequential biopsies are deemed impractical, unethical, and unsafe⁴¹. In this regard, using a PET probe to profile tumors could reduce cases of biopsy mismatch by looking at the entire tumor in an unperturbed, non-invasive setting. In the succeeding sections, the benefits and considerations are discussed using patient case studies reported in recent years as examples.

Confirmation of malignancy and antigen density

ImmunoPET may potentially provide an image-guided molecular diagnostic tool where pathological results may not be able to confirm and identify true positive disease. It detects the target antigen and quantitatively measures its expression. The imaging agent ¹⁸F-FDG has long been the standard PET tracer for detecting lesions, but it is limited to visualizing tumor metabolism. Moreover, weak tumor avidity or probe accumulation, non-specific tissue binding, and low metabolic lesions can pose problems, hindering detection⁴². Pandit-Taskar et al. conducted identification of metastatic bony lesions using the anti-PSMA PET tracer, ⁸⁹Zr-J591 and analyzed against lesions detected by ¹⁸F-FDG, bone scans (^99mTc-medronic acid (MDP)) and computed tomography (CT). ⁸⁹Zr-J591 was able to detect four occult lesions, which were undetected by FDG and other imaging assays⁴³. Out of 21 lesions, 19 were PSMA-positive as identified by ⁸⁹Zr-J591. Of these select osseous lesions, two were biopsy-proven negative, but further assessment using magnetic resonance imaging confirmed one of the lesions as metastatic with a repeat biopsy confirming the malignancy.

Optimization of pharmacokinetics

Admittedly, utility of full-length mAb as vectors of PET nuclides are limited. The long-lived blood pool circulation of these mAbs (3–8 days) yield long wait times between administration and scan acquisitions for optimum contrast between tumor and normal tissue signal. In this regard, antibody fragments are currently explored as faster alternatives. A prospective phase I/IIa clinical trial investigating ⁸⁹Zr-IAB2M, a minibody targeting PSMA in metastatic prostate cancer proved this principle (Fig. 1A-C)⁴⁴. A comparison of the pharmacokinetic properties between ⁸⁹Zr-J591 and the minibody tracer displayed faster serum clearance of the latter (Fig. 1D) with lesion accumulation at 24–48 hours (⁸⁹Zr-J591: 6–8 days after administration, Fig. 1E). Organ dosimetries were parallel for both tracers with hepatic tissue attaining the highest dose exposure.

Figure 1. Confirmation of Malignancy.

Differences in lesion detection in a metastatic prostate cancer patient using ^99mTc-MDP (bone scan) showed lesions in the ribs and vertebrae A), ¹⁸F-FDG PET scan displayed uptake in the femur and in the vertebrae B), and ⁸⁹Zr-IAB2M imaging identified more true-positive lesions than ^99mTc-MDP and ¹⁸F-FDG C). Optimization of pharmacokinetics. A comparison of serum clearance D) and lesion uptake E) between ⁸⁹Zr-IAB2M (minibody) and ⁸⁹Zr-J591 (full length mAb cognate) over time. This research was originally published at JNM Pandit-Taskar N,Donoghue JA, Ruan S, et al. First-in-Human Imaging with ⁸⁹Zr-Df- IAB2M Anti-PSMA Minibody in Patients with Metastatic Prostate Cancer: Pharmacokinetics, Biodistribution, Dosimetry, and Lesion Uptake. J Nucl Med.2016;57(12):1858–1864. © by the Society of Nuclear Medicine and Molecular Imaging, Inc. (reprinted with permission from ref.44).

Receptor occupancy and pharmacokinetics dose validation

Dose escalation studies using ⁸⁹Zr-IAB2M (anti-PSMA minibody) in patients were conducted with 10 mg, 20 mg, or 50 mg of IAB2M. Differences in biodistribution were minor across all doses. Decreased blood pool activity coupled with an increased liver and GI tract accumulation was observed over time. The highest lesion uptake was seen in the 10-mg cohort with optimal biodistribution for imaging, as well as improved delineation of bony metastatic sites. Of note, increased doses of the cold IAB2M resulted in slower serum clearance due to mass effects, although a non-significant decrease in liver uptake was noted in the 50 mg cohort. Perhaps the most impact immunoPET has contributed can be gleaned from the pioneering study investigating the biodistribution of ⁸⁹Zr-trastuzumab in patients with metastatic breast cancer (BC). Djikers et al. observed rapid hepatic excretion and low blood pool levels of the tracer in breast cancer (BC) patients who are naïve to trastuzumab with extensive HER2+ tumor mass in the liver; consequently, a false-negative readouts in distal metastatic sites was exhibited⁴⁵. The hepatic “sink” and poor uptake in metastatic lesions were attributed to slow extravasation of the drug through the vascular compartment compared to fast pharmacokinetic clearance of the mAb at low dose levels. In this study, a 10 and 50 mg loaded dose displayed terminal half-lives of 1.5 and 4.3 days respectively; in contrast, tumor penetration and accumulation of ⁸⁹Zr-trastuzumab occurred between 4–5 days. To gain perspective, administered therapeutic doses (4 mg/kg loading plus 2 mg/kg maintenance dose) reached an average terminal half-life of ~28.5 days when at steady state. Another important finding of this pivotal clinical trial was the importance of drug receptor occupancy. The fast pharmacokinetics of low trastuzumab doses led the authors to estimate drug/receptor occupancy by considering the amount of HER2 per tumor cell and the liver mass of the patient. The mass (1.2 kg) was obtained through image analysis of normalized PET/CT scans. The authors rationalized that a 50 mg dose of trastuzumab, equivalent to 2.0×10¹⁷ trastuzumab molecules (via conversion through Avogadro’s number) cannot fully saturate over a kg (1.2 kg) of tumor tissue based on the following approximations. A gram of tumor tissue is nearly comprised of ~1 × 10⁹ cells. Each single cell, on average, possesses 2 million HER2 receptor sites. Thus, in the patient’s case, there are ~2.4 × 10¹⁸ HER2 receptor molecules present in the hepatic metastases, 10-fold higher than the 50 mg dose (1.2×10³ g tumor tissue × 1×10⁹ cells/g × 2×10⁶ HER2 receptors/cell)^46–47. Majority of the dose (50 mg) accumulated in the extensive liver metastasis. This created the impetus to vary doses in patients who are naïve to trastuzumab versus those receiving this treatment with the former requiring more mAb administered (50 mg vs. 10 mg, respectively, Fig. 2).

Figure 2. Receptor occupancy.

⁸⁹Zr-trastuzumab PET biodistribution in patients given 10 mg of ⁸⁹Zr-trastuzumab (untreated) A), 50 mg ⁸⁹Zr-trastuzumab during concurrent trastuzumab treatment B), and 10 mg ⁸⁹Zr-trastuzumab during concurrent trastuzumab treatment C) show different clearance rates in the blood pool D), and should be considered when dosing patients in the clinic. This research was originally published at Clin Pharmacol Ther. Dijkers EC, Oude Munnink TH, Kosterink JG, et al. Biodistribution of ⁸⁹Zr-trastuzumab and PET imaging of HER2-positive lesions in patients with metastatic breast cancer. Clin Pharmacol Ther. 2010;87(5):586–592. (reproduced with permission from Wiley © ref.45)

Taken together, these pivotal biodistribution studies underscore the substantial dependence of mAb-based therapies (e.g. trastuzumab-emtansine (T-DM1)⁴⁸, pertuzumab⁴⁹, rituximab⁵⁰) on pharmacokinetics for personalized dosing strategies. Current clinical protocol relies on body weight to determine drug doses administered. ImmunoPET CDx can potentially transform this practice by facilitating the assessment of effective patient-tailored doses based on the extent of tumor burden and mAb pharmacokinetics.

Discordance with pathologic findings

A clinical study assessing ⁸⁹Zr-rituximab as an imaging biomarker of CD20 in patients with relapsed or refractory diffuse large B cell lymphoma was correlated against pathologic findings⁵¹. Biopsy-proven lesions (5/6 patients) showed concordance with the tumor uptake of ⁸⁹Zr-rituximab. A strong uniform staining of CD20 was correlated with a high SUV_peak of 12.8 while a moderate, heterogeneous CD20 expression corresponded to a tumor uptake of SUV_peak ~ 3.2–5.4 (Fig. 3A). In certain cases, the pathology may lead to discordance with the immunoPET data. One patient demonstrated a biopsy-mismatch with CD20 PET displaying a positive tumor uptake (SUV_peak ~ 3.8) but negative pathology (Fig. 3B). The lesion was conclusively assessed as a true positive.

Figure 3. ImmunoPET findings in relation to pathology.

Concordance A) and B) discordance of ⁸⁹Zr-rituximab-PET/CT (left) with CD20 pathology via IHC (right). Arrows point to lesions on the PET scan. This research was originally published at PLOS One Jauw YW, Zijlstra JM, de Jong D, et al. Performance of ⁸⁹Zr-Labeled-Rituximab-PET as an Imaging Biomarker to Assess CD20 Targeting: A Pilot Study in Patients with Relapsed/Refractory Diffuse Large B Cell Lymphoma. PLoS One. 2017;12(1):e0169828 and modified for use under the creative commons license https://creativecommons.org/licenses/by/4.0/).

Another concrete example was presented by Ulaner et al. investigating HER2-PET in patients with HER2-negative primary BC⁵². Of the 20 patients, 15% (3/20) were identified by ⁸⁹Zr-trastuzumab as having unsuspected HER2-positive metastases with proven pathologies. In Fig. 4, a patient who was diagnosed with ER+/HER2- invasive ductal BC presented two years later with several bone lesions and was observed HER2-PET avid. Biopsy of the right ilium (SUV~ 5.9) confirmed metastases but with an ambiguous IHC score of 2+. Confirmation of the foci as true-positive was made using MSK-IMPACT assay. Of note, the authors emphasized that the intensity of the PET tracer on foci can indiscriminately assess true- from false-positive lesions. The study reported ~30% (6/20) of the patient population was conservatively categorized as false-negative due to negative pathology even with foci avidity for the probe. The relatively high incidence of false-positive lesions was attributed to non-specific uptake of free Zr-89, particularly in osseous sites, which marginalizes the use of this nuclide for detecting bone metastases. Collectively, tumor heterogeneity can impact go/no-go treatment decisions with standard biopsy results rendering ambiguity to some extent. In these cases, immunoPET can reinforce and potentially resolve equivocal tumor pathology. However, confirmation of true-positive or -negative lesions as visualized by immunoPET needs to be meticulously validated.

Figure 4.

PET readout gave true-positive results despite discordance with biopsy findings. An ER+/HER2- invasive ductal BC patient with confirmed negative pathology in the primary lesion A) but presented with HER2-PET positive disease 2 years after primary diagnosis B). Biopsy of the same site resulted an ambiguous IHC score (2+) C). MSK-IMPACT assay confirmed the foci as truepositive D). Arrows point to lesions. This research was originally published in JNM. Ulaner GA, Hyman DM, Ross DS, et al. Detection of HER2-Positive Metastases in Patients with HER2-Negative Primary Breast Cancer Using 89Zr-Trastuzumab PET/CT. Journal of nuclear medicine : official publication, Society of Nuclear Medicine. 2016;57(10):1523–1528. ©1by the Society of Nuclear Medicine and Molecular Imaging, Inc.(reproduced with permission from Ref. 52)

Tissue distribution, radiation dosimetry and the need for sequential imaging

PET-derivatized mAbs can be utilized to determine distribution and organ exposure. This strategy has been utilized several years ago especially in the field of radioimmunotherapy. Isotopologues like ¹¹¹In/¹⁷⁷Lu⁵³, ¹¹¹In/⁹⁰Y)⁵⁴ and ^99mTc/¹⁸⁶Re⁵⁵ are ideal radionuclides as theranostics (imaging and drug agents developed in tandem)⁵⁶. Unfortunately, zirconium-89, the ideal PET nuclide for mAb radiochemistry does not have a therapeutic isotopologue nuclide^13,57 albeit zirconium has two other beta-emitting radioisotopes, ⁹⁵Zr (t_1/2 ~ 64.0 d) and ⁹⁷Zr (t_1/2 ~ 16.74 h) that could potentially be useful but have not been investigated. Despite this limitation, a number of studies continued to employ ⁸⁹Zr for tissue distribution and dosimetry studies owing to its inherent properties as a PET nuclide. This was evident in a study, which demonstrated ⁸⁹Zr-ibritumomab tiuxetan as a far better tool for evaluating distribution and dosimetry of ⁹⁰Y-labeled ibritumomab tiuxetan (Zevalin™), compared to using the single photon emission tomography (SPECT) isotope-labeled counterparts (e.g. ¹¹¹In and ¹³¹I)⁵⁸. Using SPECT can be problematic and prone to errors stemming from its lower sensitivity with low geometric efficiencies (~0.01% detected vs. emitted photons) as a consequence of its poor collimator detector response⁵⁹; thus, requiring substantial correction and compensation algorithms to achieve accurate quantification for dose-toxicity assessments⁶⁰. The study was conducted in patients with non-Hodgkins lymphoma who were given the tracer either for baseline imaging or during Zevalin treatment. Tissue binding and clearance were observed in a manner consistent with the radioimmunotherapy with a slight difference in bone uptake. Disparities in organ and whole body absorbed dose estimations are apparent when compared to SPECT-labeled ibritumomab tiuxetan. In the ⁸⁹Zr-labeled mAb, the organ receiving the highest dose was the liver (3.2 ± 1.8 mGy/MBq) followed by the spleen (2.9 ± 0.7 mGy/MBq) with a whole body absorbed dose of 0.87 mSv/MBq. Several studies utilizing ¹¹¹In-labeled ibritumomab tiuxetan showed a reversed trend, with the spleen absorbing the highest dose followed by the liver. A study by Carrasquillo et al. revealed that pharmacokinetic parameters between ¹¹¹In and ⁹⁰Y differ by ~10–15%, which can account for dosimetry differences⁶¹.

The findings from this study outline the advantages of ⁸⁹Zr-immunoPET over SPECT-based CDx to recapitulate whole body distribution of its radioimmunotherapy counterpart. Utilizing CDx further extracts dose-response correlations to determine drug efficacy and toxicities.

Predictive markers of treatment

A first-in-human study investigated by Lamberts et al. evaluated ⁸⁹Zr-MMOT0530A in pancreatic tumors and metastases expressing mesothelin (MSLN)⁶². Pre-treatment scans showed a mean SUV_max of 11.5 ± 5.6 lesions in the pancreas. Patients received the antibody-drug conjugate DMOT4039A (MMOT0530A bound to MMAE) followed by ⁸⁹Zr-MMOT0530A PET, 4 days post injection of the tracer. After treatment, 9 out of 11 patients presented with stable disease, and two patients had progressive disease. Those with progressive disease showed an uptake in liver metastasis with the PET tracer. This suggests that ⁸⁹Zr-MMOT0530A-PET can be used to visualize pancreatic cancer lesions, as well as guide individualized antibody-based treatment with the ADC DMOT4039A.

The landmark ZEPHIR study evaluated the predictive value of HER2 PET/CT in combination with FDG PET prior to T-DM1 treatment in patients with metastatic breast cancer⁶³. From the 55 patients enrolled, 16 (29%) were negative for HER2-PET while 39 patients were categorically classified as positive for HER2-PET/CT, depending on lesion heterogeneity. From the HER2-positive pool, 28 patients displayed an objective response (OR) after 3 cycles of T-DM1. In combination with post-treatment (after 1 cycle of T-DM1), a 100% positive predictive value (PPV) was achieved for HER2-PET imaging (72% PPV) in combination with early treatment FDG-PET imaging based on RECIST 1.1 (Fig. 5). Moreover, a time-to-treatment failure of ~ 11.2 months in the HER2-positive group and ~3.5 months for the HER2-negative group were identified. A negative predictive value of 88% in patients with low HER2-PET was deemed clinically significant. To date, this is the first trial that used a three-prong strategy that employed imaging biomarkers for go/no go treatment decisions in the clinic. In conclusion, these clinical trials highlighted the potential of immunoPET to measure functional effects of targeted treatment, making this imaging technique a conceivable predictive and prognostic biomarker.

Figure 5. Predictive Markers of Treatment.

Time-to-treatment failures were evaluated based on HER2-PET/CT A), early FDG-PET/CT B) and combination of both HER2- and FDG-PET/CT C). This research was originally published in Ann Oncol. Gebhart G, Lamberts LE, Wimana Z, et al. Molecular imaging as a tool to investigate heterogeneity of advanced HER2-positive breast cancer and to predict patient outcome under trastuzumab emtansine (T-DM1): the ZEPHIR trial. Ann Oncol. 2016;27(4):619–624. (Reproduced with permission From Oxford University Press Ref. 63).

Practical Considerations

While immunoPET CDx may seem straightforward, several aspects of using this imaging technique need to be deliberated. The amount of dose administered and the interval between tracer administration and imaging acquisition warrant investigation to obtain an optimized contrast between lesions and background. In the case of ⁸⁹Zr-trastuzumab, the optimal imaging time for a ~37 MBq (50 mg) intravenous injection was observed between 4–5 days after injection^45,65. At this period, low blood pool activity and high tumor avidity was established. Imaging at longer periods >6 days can compromise the spatial resolution and image quality⁴⁵. At higher activities (~185 MBq/50 mg) administered, ⁸⁹Zr-trastuzumab still generated high quality spatial resolution in images acquired between 5–6 days post-injection^{45,51–52,64–65}. The scan periods of 4–6 days depending on the dose are typical for other full-length mAb tracers in clinical trials. For smaller biologics-based tracers, ⁸⁹Zr-IAB2M, for example, demonstrated shorter interval wait times with the best lesion to background ratio identified at 48 h p.i.⁴⁴ Safety profiles of ⁸⁹Zr-labeled mAbs require careful assessment to limit radiation-related toxicities. Whole body effective doses reported in a number of early phase studies ranged from 0.41 mSv/MBq for ⁸⁹Zr-IAB2M⁴⁵, 0.87 ± 0.14 mSv/MBq for ⁸⁹Zr-ibritumomab tiuxetan⁵⁸, 0.47 mSv/MBq for ⁸⁹Zr-trastuzumab⁶⁶ and 0.264 mSv/MBq for ⁸⁹Zr-panitumumab⁶⁷ whereas FDG-PET⁶⁸ had a reported mean effective dose of 0.0199 ± 0.0032 mSv/MBq.

Engagement of immunoPET CDx as predictive imaging biomarkers in the clinic should continuously be explored to support clinical translational efforts towards precision medicine. It has already shown beneficial in accurately profiling lesions at the molecular level when pathology is incorrect, discovering the density of targets available, and determining the biodistribution of therapy before treating the patient. Sequential imaging in test-retest studies can provide a viable tool to appropriately dose patients, but should be used with caution during treatment regimens. In a nutshell, immunoPET is still at its early stages of clinical development and will most likely require further standardization (i.e. streamlined SUV readout analysis, chemistry optimization) and validation through other molecular profiling tools. Once harnessed, its benefits can provide a powerful impact in patient management.