Prospective Characterization of Circulating Tumor Cell Kinetics in Patients With Oligometastatic Disease Receiving Definitive Intent Radiation Therapy

Abstract

PURPOSE

There are currently no predictive molecular biomarkers to identify patients with oligometastatic disease (OMD) who will benefit from definitive-intent radiation therapy (RT). We prospectively characterized circulating tumor cell (CTC) kinetics in patients with OMD undergoing definitive-intent RT.

METHODS

This prospective correlative biomarker study included patients with any solid malignancy ≤5 metastatic sites in ≤3 anatomic organ systems undergoing definitive-intent RT to all disease sites. Circulating tumor cells (CTCs) were captured and enumerated using a biomimetic cell rolling and nanotechnology-based assay functionalized with antibodies against epithelial cell adhesion molecule, against human epidermal growth factor receptor 2, and against epidermal growth factor receptor before and during RT and at follow-up visits up to 2 years post-RT.

RESULTS

We enrolled 43 patients with a median follow-up of 14.3 months. The pretreatment CTC level (cells captured/mL) was not associated with the number of disease sites (median one metastatic site/patient, range 1-5) or metastasis location (bone, brain, visceral) on Wilcoxon signed-rank test, P > .05. Post-RT, 56% of patients received systemic therapy, and 72% of patients experienced subsequent local or systemic progression. For 90% of patients, a CTC level <15 within 130 days post-RT corresponded to a durable control of irradiated lesions. Patients with a favorable versus an unfavorable clearance profile experienced significantly longer progression-free survival after RT (median 13 v 4 months, log-rank test, P = .0011). On logistic regression, CTC level >15 at a given time point was associated with clinical disease progression within the subsequent 6 months (odds ratio 3.31, P = .007). In 26% of patients with disease progression, a CTC level >15 preceded radiographic or clinical progression.

CONCLUSION

CTCs may serve as a biomarker for disease control in OMD and may predict disease progression before standard assessments for patients receiving diverse cancer-directed therapies.

INTRODUCTION

Patients with oligometastatic disease (OMD), an intermediate state between localized and wide-spread metastatic disease, may experience a progression-free survival (PFS) or overall survival (OS) benefit from aggressive metastasis-directed therapy.^1-5 Guidelines define OMD as ≤5 metastatic lesions on imaging, an optional controlled primary tumor, where all metastatic sites are safely treatable on the basis of expert consensus given the absence of rigorously validated criteria.² Thus, this definition encompasses true OMD and occult polymetastatic disease which is treated per a vastly different paradigm. There are presently no clinically available assays to assist in defining OMD, selecting appropriate candidates for OMD-directed therapy, nor identifying patients with persistent disease requiring prompt initiation of systemic therapy, highlighting an urgent need for biomarkers for management of OMD.

CONTEXT

• Key Objective

• There are currently no predictive molecular biomarkers to identify patients with oligometastatic disease (OMD) who will benefit from definitive-intent radiation therapy (RT). We prospectively characterized circulating tumor cell (CTC) kinetics in patients with OMD undergoing definitive-intent RT to all disease sites.

• Knowledge Generated

• Patients with a favorable versus an unfavorable clearance profile experienced significantly longer progression-free survival after RT. On logistic regression, elevations in CTC levels at a given time point were associated with clinical disease progression within the subsequent 6 months including a subset of cases where elevation in CTC level preceded radiographic or clinical progression.

• Relevance

• CTCs may serve as a biomarker for disease control in OMD and have potential to further define the role and timing of RT and systemic therapy for OMD by predicting disease progression prior to standard assessments for patients receiving diverse cancer-directed therapies.

Circulating biomarkers including cell-free DNA (cfDNA) and circulating tumor cells (CTCs) show potential as liquid biopsies to diagnose cancer, monitor response, detect residual disease, and recognize recurrence before standard surveillance practices.⁶ Although cfDNA has substantial emerging indications, it is challenging to apply in the setting of OMD given the wide array of histologies under investigation and the ability to only detect a priori specified mutations that may not reflect clones surviving therapy. In contrast, sensitive CTC capture assays are capable of detecting CTCs from a wide range of histologies and mutational profiles in the peripheral circulation.^7,8

We performed a prospective correlative biomarker study inclusive of patients with OMD receiving definitive-intent radiation therapy (RT) to all disease sites in which we characterize CTC profiles associated with treatment outcomes.

METHODS

Patients and Biospecimens

This prospective single-institution correlative biomarker study included patients with any solid malignancy and OMD defined as ≤5 metastatic sites in ≤3 anatomic organ systems receiving definitive-intent RT to all active disease sites. Patients with de novo oligometastatic, oligorecurrent, oligoprogressive, and oligopersistent disease meeting the above criteria were included.¹ Subsequent lines of therapy were given per treating physician's discretion. Whole blood was collected in sodium heparin tubes. Biospecimen collections occurred before RT (baseline), during RT, and at standard-of-care follow-up visits up to 24 months post-RT. Enrollment was from October 5, 2015, to October 28, 2020. Providers were blinded to CTC enumeration results.

Ethics Approval

All appropriate ethics approvals and consents were obtained from the University of North Carolina Institutional Review Board.

CTC Enumeration

In brief, CTCs were captured and enumerated using a previously reported biomimetic cell rolling and nanotechnology-based assay. The capture surfaces were functionalized with three antibodies against epithelial cell adhesion molecule (aEPCAM), against human epidermal growth factor receptor 2 (aHER2), and against epidermal growth factor receptor (aEGFR) on poly(amidoamine) (PAMAM) dendrimers-coated surface for strong multivalent binding, along with recombinant human E-selectin that induces cell rolling.⁷ The surfaces after capture were imaged after immunocytochemistry for pan-cytokeratin (cytokeratin [CK]), leukocyte marker CD45, and 4’,6-diamidino-2-phenylindole (DAPI) nuclear stain. CTC levels were determined from DAPI+/CK+/CD45– cells and expressed per mL of whole blood.

CTC Capture Surface Preparation

Capture surfaces were fabricated as previously described.⁷ Generation 7 PAMAM dendrimer (Dendritech, Midland, MI) was purified by centrifugal filtration, lyophilized, and partially carboxylated by reaction with succinic anhydride (Sigma-Aldrich, St Louis, MO) in dimethyl sulfoxide at a molar ratio of 358:1, resulting in approximately 70% of terminal amines converted to carboxyl groups. Epoxy-functionalized glass slides (Tekdon, Myakka City, FL) were treated overnight with 5,000 MW and heterobifunctional amine-poly(ethylene glycol)-carboxymethyl polymers (JenKem, Plano, TX) at 0.5 mg mL⁻¹ in distilled deionized (DDI) water. The polyethylene glycol was activated with 15 mM 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide hydrochloride (EDC, Sigma-Aldrich) and 25 mM N-hydroxysuccinimide (NHS, Sigma-Aldrich) in DDI water for 30 minutes. PAMAM-COOH was added at 0.1 mg mL⁻¹ in sodium phosphate buffer, pH 8 with 150 mM sodium chloride and incubated overnight. The surfaces were activated once more with EDC and NHS for 30 minutes before treatment for 60 minutes with a mixture of mouse anti-human monoclonal antibodies (R&D Systems, Minneapolis, MN) each at 1 μg mL⁻¹ in sodium phosphate, pH 8 with 150 mM sodium chloride. Mouse anti-human monoclonal antibodies consisted of aEPCAM (R&D Systems AF960), aHER2 (R&D Systems AF1129), and aEGFR (R&D Systems AF231).⁹ Finally, surfaces were incubated with 1 μg mL⁻¹ recombinant human e-selectin Fc chimera protein (R&D Systems 724-ES) in phosphate buffered saline (PBS) for 60 minutes. Slides were stored at 4°C in PBS for up to 4 weeks before use.

CTC Purification and Quantification

Blood tubes were gently agitated, and 6 mL whole blood was separated by Ficoll-Paque PLUS (GE Healthcare, Uppsala, Sweden). The buffy coat was purified twice by centrifugation to isolate mononuclear cells and resuspended in 0.1 mL Dulbecco's Minimum Essential Medium (Corning Life Sciences, Tewksbury, MA). The cell suspension was pumped through a flow chamber⁸ containing the capture slide. The flow chamber consisted of a pair of 55 mm long, 5 mm wide, and 0.15 mm tall channels connected by a length of 0.16ʺ inner diameter tubing. The flow rate was controlled by a NE-1000 syringe pump (New Era, Farmingdale, NY) at 0.025 mL/min for 20 minutes, followed by 0.090 mL/minutes for 15 minutes, followed by a rinse in the reverse direction with PBS at 0.090 mL/minutes for 15 minutes.

Slides were removed from the chamber, fixed with 4% paraformaldehyde (Polysciences, Warrington, PA), permeabilized with 0.2% Triton X-100 (Sigma-Aldrich), and treated with 2 wt.% bovine serum albumin (BSA, Sigma-Aldrich) to reduce nonspecific binding. Cells were then sequentially stained with the following antibodies: rabbit anti-human pan CK (1:50, Abcam 9377), goat anti-human AlexaFluor 568 or AlexaFluor 647 conjugated secondary antibody (1:500, Thermo Fisher Scientific, Waltham, MA), mouse anti-human CD45 (1:500, BD Biosciences 555480, San Jose, CA), and goat anti-mouse AlexaFluor 488 conjugated secondary antibody (1:500, Thermo Fisher Scientific). Nuclei were labeled with DAPI (Thermo Fisher Scientific), and slides were sealed with mounting media under a cover glass.

Slide surfaces were imaged with a Zeiss 701 confocal laser scanning microscope with 405-nm diode UV laser, 488-nm mutliline Ar laser, and 561-nm solid state diode laser at 20×/0.8 Plan-Apocromat objective (Carl Zeiss, Munich, Germany) or Zeiss Axio Observer wide field microscope with Colibri 7 LED light source with a 90 HE filter set and either a 20×/0.5 or 10×/0.3 Plan-Neofluar objective and Axiocam 503 mono CCD Camera (Carl Zeiss). A motorized stage on both microscopes enabled collection of the entire slide surface in tiled images. After color balancing, each color channel was exported as a jpeg and analyzed using a custom algorithm in Image J (NIH). The algorithm identified particles in the cytokeratin channel with brightness and circularity similar to a cell. Any CK-positive particles with a minimum signal in the CD45 channel under or immediately around the particle were discarded. Finally, particles were identified in the DAPI using brightness and circularity thresholds. Particles were identified as CTCs if the centroid of a DAPI particle was within half the radius of the centroid of a CK particle.

End Points and Statistical Analysis

Disease status was assessed per standardized RECIST 1.1 criteria at each time point with available CTC level and clinical evaluation/imaging obtained per standard of care. Clinical documentation, radiology reports, and primary images were reviewed by two clinicians to determine whether target lesions met criteria of complete response, progressive disease, or noncomplete response/nonprogressive disease. On exploratory analysis, disease status was correlated with CTCs as a continuous and ordinal variable (cut point upper bound of the third quartile). Regarding kinetics, a favorable (v unfavorable) CTC clearance profile was defined as a decrease (v increase) in CTCs between pretreatment and the end of treatment. PFS was estimated from the end date of RT to the date of the event. Data were analyzed with a generalized linear mixed model. A two-sided P value of <.05 was considered significant. Statistical analyses were performed with R version 4.0.2.¹⁰

RESULTS

We enrolled 43 patients with a median follow-up of 14.3 months post-RT (range, 1.4-24.7 months) corresponding to 255 CTC measurements. Samples from 16 time points, 6% of samples, were excluded from analysis because of technical issues including errors during sample processing, shipping delays, sample coagulation, and insufficient biospecimen. Patient characteristics are summarized in Table 1. The median number of distinct metastatic lesions per patient was 1 (range, 1-5). Most patients received stereotactic body RT (SBRT; 72%) versus conventional or hypofractionated courses of RT (28%).

TABLE 1.

Patient Characteristics

At baseline, all patients had detectable CTCs (median baseline, 28; range, 0.17-1,085). There was no significant association between the pretreatment CTC level and presence of one versus ≥2 metastatic sites (median CTC 22 v 42, Wilcoxon signed-rank test, P = .37), Figure 1. The location of the metastatic site was not associated with the CTC level when categorized as bone, brain, or visceral (median CTC 36 versus 5 v 28, Wilcoxon signed-rank test, P > .05 for all comparisons).

FIG 1.

Baseline CTC levels and association with disease state. All patients had detectable CTCs before radiation therapy. There was no association between pretreatment CTC count and the total number of active oligometastatic sites. There was also no association between the pretreatment CTC count and the location of metastatic sites with regards to bone, brain, and visceral metastatic disease. CTC, circulating tumor cell.

During RT, the median CTC level declined: baseline 22 (range, 0.17-99; SD, 30), end of RT 11 (range, 0.00-242; SD, 58), and first post-treatment follow-up within 130 days post-RT median 11 (range, 0.00-157; SD, 37) for 17 patients evaluable at all time points.

Therapy after RT was determined per physician discretion independent from CTC levels. After RT, 72% of patients experienced disease progression (n = 31); 3% (n = 1) in the irradiated lesion alone and 33% in the irradiated lesion as well as systemic disease (n = 10), and 64% experienced systemic progression alone (n = 20). For 90% of patients (n = 27 of 30), CTC ≤15 within 130 days post-RT corresponded to long-term local control of irradiated lesions. Among patients with disease progression, 26% experienced an increase in CTC >15 before radiographic or clinical detection of disease progression (n = 8 of 31). Patients with a favorable versus unfavorable clearance profile experienced significantly longer PFS post-RT (median 13 v 4 months, log-rank test, P = .0011; Fig 2).

FIG 2.

CTC kinetics during RT and association with PFS. During treatment, a favorable kinetic profile defined as a decrease in CTC count between pretreatment and end of treatment versus an unfavorable profile defined as the opposite was associated with a durable response to therapy as evidenced by the difference in PFS between patients with the two CTC profiles. Median PFS for patients with a favorable clearance profile was 13 months compared with 4 months for those with an unfavorable clearance profile. CTC, circulating tumor cell; PFS, progression-free survival; RT, radiation therapy.

At each follow-up time point, we evaluated whether patients experienced disease progression allowing for multiple progression events per patient during their disease course. There were 43 progression events versus 109 nonprogression events in total. On logistic regression, CTC >15 at a given time point was associated with clinical disease progression within the subsequent 6 months (odds ratio, 3.31; P = .007). During this period, 56% (n = 24) patients received systemic therapy (cytotoxic chemotherapy, hormone therapy, immunotherapy, kinase inhibitors) either immediately after RT or at the time of progression. This association remained when controlling for prior CTC levels and type of systemic therapy.

DISCUSSION

Recent randomized clinical trials of SBRT for OMD-directed therapy show favorable outcomes for a subset of patients.^3-5 Prognostic and predictive biomarkers are needed to identify patients with OMD versus occult polymetastatic disease who will benefit from OMD-directed therapy.

In this correlative biomarker study, we characterize CTC profiles associated with favorable response versus disease progression in a prospective cohort of patients receiving definitive-intent RT for OMD. These data demonstrate that optimized CTC enumeration technologies can detect CTCs in the setting of OMD, and furthermore, CTC profiles during cancer-directed therapies are strongly associated with, and potentially predictive of, therapeutic response. The ability to distinguish patients with response versus nonresponse to OMD-directed therapy is critical for identifying patients for whom prompt initiation of systemic therapy is indicated.

These findings are consistent with studies of widely metastatic breast, colorectal, and castrate-resistant prostate cancers wherein CTC enumeration is prognostic of PFS or OS using the CellSearch platform.^11-13 However, the utility of existing commercial platforms in predictive and low-disease burden settings is limited in part due to assay sensitivity and specificity. Here, we used a second-generation assay with biomimetic rolling and antibodies targeted to various antigens expressed on CTCs.^7,14,15

Our study has important limitations. First, the sample size is modest and includes various histologies. Second, the RT dose and fractionation were not standardized. Third, after RT, additional lines of therapy and follow-up intervals were variable. Bearing these limitations, our findings show potential for CTC profiles to predict response to OMD-directed RT and subsequent recurrence.

In conclusion, CTCs may serve as a biomarker for disease control in OMD and have potential to further define the role of both RT and systemic therapy for OMD.

DATA SHARING STATEMENT

Research data are stored in an institutional repository and will be shared on request to the corresponding author.

AUTHOR CONTRIBUTIONS

Conception and design: Shivani Sud, Jacob Hall, Jiyoon Bu, Ja Hye Myung, Andrew Z. Wang, Seungpyo Hong, Dana L. Casey

Financial support: Andrew Z. Wang, Seungpyo Hong, Dana L. Casey

Provision of study materials or patients: Andrew Z. Wang, Seungpyo Hong, Dana L. Casey

Collection and assembly of data: Shivani Sud, Michael J. Poellmann, Jacob Hall, Jiyoon Bu, Ja Hye Myung, Andrew Z. Wang, Dana L. Casey

Data analysis and interpretation: Shivani Sud, Michael J. Poellmann, Jacob Hall, Xianming Tan, Andrew Z. Wang, Seungpyo Hong, Dana L. Casey

Manuscript writing: All authors

Final approval of manuscript: All authors

Accountable for all aspects of the work: All authors

AUTHORS' DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST

The following represents disclosure information provided by authors of this manuscript. All relationships are considered compensated unless otherwise noted. Relationships are self-held unless noted. I = Immediate Family Member, Inst = My Institution. Relationships may not relate to the subject matter of this manuscript. For more information about ASCO's conflict of interest policy, please refer to www.asco.org/rwc or ascopubs.org/po/author-center.

Open Payments is a public database containing information reported by companies about payments made to US-licensed physicians (Open Payments).