Radioimmunotherapy for peritoneal carcinomatosis: Preclinical proof of concept to clinical translation

Summary

Peritoneal carcinomatosis (PC), characterized by the dissemination of metastatic tumor cells throughout the peritoneal cavity from several gastrointestinal and gynecological malignancies, has significantly compromised patient survival. The standard of care is cytoreductive surgery with or without intraperitoneal chemotherapy. However, surgical resection often leaves behind microscopic or clinically occult disease due to the complex anatomy of the peritoneum, where intraperitoneal chemotherapy and systemic chemotherapy have shown limited success. To improve the therapeutic outcome, targeted therapy using radionuclides such as alpha, beta, and Auger emitters delivered by antibodies is actively being investigated. While preclinical murine models of PC have shown the potential of radioimmunotherapy (RIT) using various radioisotopes across a wide spectrum of antigen targets and tumor diagnoses with acceptable toxicities, successful clinical trials are lacking. Here, we retrospectively summarize preclinical and clinical PC studies, consider their translational potential, and examine paths to development that maximize the clinical benefit of RIT in this context.

Graphical abstract

Aguirre et al. provides an in-depth summary of antibody-based radioimmunotherapy of peritoneal carcinomatosis in preclinical models and patients. Because of its prevalence, limited treatment options, and dismal survival statistics, PC cure remains a major unmet need. Widening the therapeutic window of RIT in PC should accelerate its clinical translation.

Introduction

Peritoneal carcinomatosis (PC) is a metastatic disease, characterized by the dissemination of metastatic tumor cells throughout the peritoneal cavity. Nearly all malignancies, particularly carcinomas of the stomach, colon, small intestine, appendix, gallbladder, pancreas, ovary, breast, and uterus, have the potential to disseminate and grow in the peritoneal cavity.^1,2,3 10%–50% of patients with recurrent gastrointestinal cancers have tumor recurrence confined to the peritoneal cavity and ultimately die from complications of locoregional tumoral spread.¹ PC has long been regarded as an incurable condition due to its difficulty to treat surgically and its perceived poor response to chemotherapy.⁴ Because of this, PC is highly morbid, affecting >70,000 patients in the US alone with a 25% mortality rate.⁵

The primary treatment option for PC is systemic therapy with cytoreductive surgery (CRS) with or without intraperitoneal chemotherapy (IPC), used for selected patients with lower-volume, more indolent, or chemoresponsive disease.⁶ In general, long-term survivors typically have had grossly complete CRS.^4,7 However, surgical resection often leaves behind microscopic or not clinically apparent lesions due to the complex anatomy of the peritoneum; systemic chemotherapy or IPC has been successfully employed for disease control, though IPC has only been proven effective in ovarian cancer trials.⁸

For decades, the addition of IPC, which involves direct infusion of chemotherapy heated to 41°C–43°C into the peritoneal cavity, was thought to enhance cytotoxicity when compared with conventional chemotherapy for PC from colon, gastric, pancreatic, and ovarian cancers. However, a randomized trial (PRODIGE 7) recently demonstrated no improvement in median overall survival (OS) regardless of CRS + IPC, IPC plus systemic chemotherapy, or chemotherapy following optimal CRS.⁹ The numerous treatment-related complications including hematologic, gastrointestinal, and neurological toxicities, as well as bowel fistulas, anastomotic disruptions, bile leakage, or pancreatitis could also negate any potential benefits from IPC.¹⁰ In contrast to the median survival of 42 months after complete CRS, those with incomplete CRS lived only 14 months. Survival depends on the histology of disease, complications from surgery, and the need for reoperation.⁷ Because of these unfavorable results, developing novel therapies for PC remains a major unmet need.

Given the importance of CRS, finding residual and occult tumor is key. Yet, noninvasive imagings commonly used to stage and stratify patients with cancer are generally insensitive or nonspecific in patients with PC. Imaging glucose metabolism with fluorodeoxyglucose- positron emission tomography (FDG PET) is not useful in assessing resectability or PC index (PCI).¹¹ In addition, it correlates poorly with the PCI, which is the standard of practice for measuring the burden of disease at the time of surgery. A high PCI (>20) is a relative contraindication to cytoreduction except for very low-grade cancer, such as well-differentiated appendiceal cancer (AC). In one study, 16% of patients who were thought to be appropriate candidates for cytoreduction were found to have a high burden of disease, prompting immediate abdominal closure following laparotomy.¹² In fact, both PET/computed tomography (CT) and CT are poor predictors of disease with PCI > 20; PET/CT added no additional benefit when compared to CT alone in the majority (64%–100%) of patients.¹³

Radioimmunotherapy

Definitive management of locoregionally advanced PC presents a major challenge and often consists of a combination of surgical, intraperitoneal, and systemic therapy approaches. Adjuvant radiation has been paramount in prolonging progression-free survival post operatively.¹⁴ However, this method is confined to solid tumors in discoverable locations, thus greatly limiting its feasibility for effective use in many metastatic cancers. Radiopharmaceutical therapy (RPT; aka, endoradiotherapy) delivers radiation to microscopic and occult disease that external beam cannot.¹⁵ Durable remission using RPT has remained elusive. Despite high antigen specificity, intravenous (i.v.) use of radiolabeled monoclonal antibodies (mAbs) often does not result in high-contrast PC tumor targeting. The critical hurdle—that is, low overall tumor uptake and high normal-tissue background with radiolabeled IgG mAbs—has not been adequately addressed.¹⁶ Two independent approaches to improve the therapeutic window (maximum tolerated dose [MTD] vs. minimum effective dose [MED]) are (1) compartmental RPT by direct injection into the body cavity (e.g., peritoneum) and (2) pretargeted or multistep RPT using novel bispecific antibody (BsAb) constructs and compatible small radioligands.¹⁷

Tumor-specific mAbs have been explored for imaging and therapy of PC.¹⁸ Attractive targets are carcinoembryonic antigen (CEA) or glycoprotein A33 (GPA33), antigens in gastrointestinal (GI) cancers like colorectal cancer; human epidermal growth factor Receptor 2 (HER2) is expressed on ovarian and other genitourinary malignancies. These antibodies can be armed with fit-for-purpose radionuclides that are suitable for diagnostic, therapeutic, or theranostic (therapeutic plus diagnostic) properties. Cytotoxic isotopes that emit beta (β⁻), alpha (α), or Auger-Meitner (e⁻) radiation can irradiate a defined tumor volume. Short path length, high LET (linear energy transfer) alpha particles cause unrepairable DNA double-strand breaks in an oxygen-independent manner within precise tumor margins, which is ideal for small volume, microscopic PC.¹⁹ On the other hand, low LET beta emitters with a longer path length in tissue are more suitable for bulky solid tumors by their cross-firing of near neighbors.²⁰ Herein, we review the last decade of radioimmunotherapy (RIT), both i.v. and intraperitoneally (i.p.), directed at PC (Figure 1; Table 1).

Figure 1.

A schematic of RIT and PRIT

(A) A diagram of classic radioimmunotherapy (RIT). RIT is usually done in one step in which a radioisotope is directly conjugated to a monoclonal antibody before being administered as a single drug.

(B) Diagram of pretargeted radioimmunotherapy (PRIT) method. The first step shown in blue is pretargeting of a multispecific antibody. Once the unbound fraction in blood reaches a safe level, the payload with strong avidity for the pretargeted antibody (shown in red) is administered in a second step (e.g., chelated radioisotope). The unbound payload is designed to clear the body rapidly thereby creating high therapeutic indices for bystander normal tissues, thus widening the therapeutic window. To reduce the unbound antibody to a safe level, a clearing agent (shown in green) has been used successfully in some systems. Created in https://BioRender.com.

Table 1.

Summary of preclinical testing of RPT using i.p. tumor models

The field lacks a harmonized metric by which to judge therapeutic efficacy and response in these models. This speaks to the need to standardize response and toxicity criteria as the field moves forward.

Preclinical RIT in PC

Beta-RIT in PC

Beta therapies are the only therapies to have successfully made it to human treatments and remain more commonly used in clinical trials.²¹ Thus, they remain an attractive option for preclinical modeling.

In clinically relevant models, the addition of RIT using ¹⁷⁷Lu (t_1/2 = 6.7 days), a beta emitter following CRS, was compared with hyperthermic intraperitoneal chemotherapy (HIPEC) following CRS. Wag/Rij rats were directly injected i.p. with syngeneic rat colon carcinoma CC531 to establish advanced PC, defined as macroscopic lesions (>3 mm), such that disease could be surgically quantified and resected. Rats were treated with i.p. RIT (74 MBq, 150 MBq/kg) using ¹⁷⁷Lu-MG1, a mouse mAb raised against CC531. HIPEC was performed by a standard closed abdominal perfusion technique using mitomycin-C. By day 5, the CRS + HIPEC group showed a decrease in anastomotic and abdominal wall wound strength when compared to the CRS + RIT group.²² Using the same tumor model in an efficacy study, CRS + HIPEC was compared to CRS + RIT or CRS alone. Rats treated with CRS + HIPEC were found to be lethargic, developing diarrhea, followed by significant weight loss. Survival after CRS was significantly improved by ¹⁷⁷Lu RIT, while adjuvant HIPEC did not improve survival.²³ Beyond well-studied colorectal, ovarian, and gastric cancers, other solid tumors that metastasize to the peritoneum though less prevalent are highly lethal. Clear cell renal cell carcinoma (ccRCC) can disseminate to the peritoneum. Beta emitting, ¹⁷⁷Lu-labeled G250, an anti-carbonic anhydrase IX antibody, was administered i.v. in mice xenografted with i.p. SK-RC-52 ccRCC. Mice receiving 13 MBq ¹⁷⁷Lu-1,4,7,10-tetraazacyclododecane tetraacetic acid (DOTA)-G250 i.v., versus ¹⁷⁷Lu-MOPC21, a nonspecific mAb, and a non-treatment control, showed significantly prolonged survival with minimal toxicities.²⁴

To study the effect of i.p. RIT using ⁶⁴Cu (t_1/2 = 12.7 h, a positron and beta emitting isotope) in a minimal i.p. disease model, the human colorectal cancer (CRC) cell line HCT116 xenografted i.p. in athymic nude mice was treated i.p. with 22.2 MBq ⁶⁴Cu-PCTA (chelator: p-SCN-Bn-PCTA)-cetuximab (1,000 MBq/kg body weight), or 22.2 MBq ⁶⁴Cu-PCTA-trastuzumab, cetuximab without radioisotope, or trastuzumab without radioisotope. Trastuzumab was the negative control given its low binding to the HER2-negative CRC cell line. Significant differences in survival were noted between ⁶⁴Cu-PCTA-cetuximab and other treatment groups, without myelotoxicity or weight loss. These results demonstrated that the treatment of PC by i.p. injection using ⁶⁴Cu-PCTA-cetuximab could be effective.²⁵

⁶⁴Cu was further tested in RIT using the commonly used cetuximab targeting HER1 in a disseminated peritoneal model of gastric cancer. This study evaluated the utility of a histone deacetylase inhibitor, vorinostat, which has been shown to sensitize gastrointestinal cancer to external radiation. Treatment of NUGC4 tumor-bearing mice with i.p. ⁶⁴Cu-cetuximab alone, ⁶⁴Cu-cetuximab plus vorinostat, or vorinostat alone was compared. It was found that coadministration of ⁶⁴Cu-cetuximab with vorinostat greatly increased survival compared to ⁶⁴Cu-cetuximab alone, while ⁶⁴Cu-cetuximab also prolonged survival compared to no-treatment controls. No significant toxicity was seen in any of the treatment groups, suggesting that combination treatment irrespective of dosing regimen did not enhance normal tissue radiosensitivity.²⁶

⁶⁴Cu was again tested in another HER1-expressing cancer, pancreatic cancer, which has the capacity for peritoneal dissemination with poor prognosis. Adjuvant i.p. anti-HER1 RIT using ⁶⁴Cu on post-operative day 1 after CRS of an i.p. xPA-1-DC pancreatic tumor model was compared to the standard of care, gemcitabine alone, or no treatment. Progression-free survival and OS were significantly improved with minimal toxicity when compared to no-treatment controls, although there was no significant difference when compared to gemcitabine alone.²⁷

Beta-RIT in PC remains a promising design for treatment of this metastatic disease. In many studies with various cancer types and targets, prolonged survival was seen compared to no-treatment controls; however, in some cases, little improvement was seen against current methods alone.

Alpha-RIT in PC

Alpha particle therapy has great potential for small-volume disease. The last decade has witnessed an explosion of interest in alpha emitters with a range of half-lives from minutes (bismuth 213, 46 min), to hours (astatine 211, 7.2 h; lead 212, 10.6 h), to days/weeks (actinium 225, 10 days; thorium 227, 18.7 days), each with unique use cases in PC.

In a murine study, the relationship of dose and toxicities was investigated comparing ²¹¹At (t_1/2 = 7.21 h), ²²⁷Th (t_1/2 = 18.7 days), ²¹³Bi (t_1/2 = 45.6 min), and ²¹²Pb (t_1/2 = 12.7 h), all alpha-emitting radioisotopes in an i.p. tumor model. Dose escalation with i.p. doses of 0.74 (20 μCi), 1.11 (30 μCi), and 1.48 MBq (40 μCi) for both ²¹¹At- or ²²⁷Th-labeled trastuzumab was tested against LS-174T, a HER2-positive CRC cell line. Toxicity was observed at doses >40 μCi, evidenced by greater weight loss and decreased platelet count. After adjustment for decay half-life and energy, 1.11 MBq (30 μCi) treatment dose of ²¹³Bi and ²¹²Pb was used for treatment comparison. While there was survival benefit compared with treatment control, there was no statistical difference in median survival between these alpha emitters. Milenic et al. concluded that the choice of radioisotope for α-RIT in the i.p. tumor model can be based on radioisotope half-life, physical parameters, and product availability.²⁸

In a similar study, Derrien et al. examined the effects post CRS in ovarian cancer, where alpha-RIT was compared with HIPEC in a murine model of postoperative ovarian PC, and SHIN-3 tumor-bearing mice were treated by HIPEC using cisplatin, 7.4 MBq (300 MBq/kg) or 11.1 MBq (444 MBq/kg) of i.p. ²¹³Bi-B-B4, an anti-CD138 mAb, or a combination of HIPEC and alpha-RIT, or no treatment. Alpha-RIT significantly improved survival, whereas HIPEC alone or with alpha-RIT was not effective, partly due to toxicity.²⁹

Similarly, Deshayes et al. used a small-volume peritoneal tumor model, with tumors <100 mm³, to mimic the disease state after CRS. Mice bearing i.p. AN3CA Mullerian-inhibiting substance receptor type II (MISRII)-positive ovarian cancer were treated with alpha-RIT or beta-RIT using the anti-MISRII mAb 16F12. Mice were treated with i.v. 16F12 carrying (1) 1,480 MBq/kg of ²¹³Bi or (2) 10 MBq (400 MBq/kg) of ¹⁷⁷Lu, or i.p. 16F12 carrying (3) 12.9 MBq (516 MBq/kg) of ²¹³Bi, or (4) 50 MBq (2,000 MBq/kg) of ¹⁷⁷Lu. i.p. RIT with ¹⁷⁷Lu-16F12 was more effective in delaying tumor growth than ²¹³Bi-16F12. High tumor-to-blood radioactivity ratio was observed with either nuclide given i.p.; this is thought to be in part due to the need for slow clearance from peritoneum into blood when compared to i.v. injection. Hematological toxicity, marked by transient decrease in hemoglobin, was more pronounced with ¹⁷⁷Lu than ²¹³Bi therapy; weight loss was noted only in those treated with ¹⁷⁷Lu. Radiolabeled 16F12 could offer another therapeutic area to pursue, specifically for i.p..³⁰

Kasten et al. also examined alpha IP-RIT without CRS in ovarian cancer using ²¹²Pb-376.96, a murine mAb specific for human B7-H3 (CD276) expressed on ovarian cancer cells combined with an in vivo alpha-emitter generator (t_1/2 = 10.6 h). Using an ovarian i.p. model (ES-2 or A2780cp20), mice were treated with 0.17–0.7 MBq of i.p. ²¹²Pb-376.96 or the irrelevant isotype-matched ²¹²Pb-F3-C25. Mice treated with ²¹²Pb-376.96 survived 2–3 times longer than mice treated with ²¹²Pb-F3-C25 or non-treated controls; transient weight loss was noted in the treated groups. These results support additional RIT studies with ²¹²Pb-376.96 for potential to guide clinical development.³¹

Epidermal growth factor receptor (EGFR), also known as HER1/ErbB1, is an attractive target in many cancer types. Milenic and Brechbiel used the anti-EGFR mAb cetuximab to deliver ²¹²Pb to EGFR-positive i.p. CRC cancer model LS174T. In a series of studies, a 0.37 MBq (10 μCi) dose of ²¹²Pb-labeled cetuximab given i.p. has no toxicity as measured by animal weight loss and hematologic markers; the therapeutic efficacy was similar to that of i.p. ²¹²Pb-trastuzumab. Concurrent i.p. treatment with both ²¹²Pb-cetuximab and ²¹²Pb-trastuzumab provided additional benefit that was greater than either radiolabeled antibody alone. When combined with gemcitabine given prior to α-RIT, median survival was significantly prolonged, while combination with carboplatin reduced the effectiveness. This difference was likely due to gemcitabine being a more efficient radiosensitizer than carboplatin.³² These results suggest that while ²¹²Pb-cetuximab has potential for treating HER1-expressing tumors, additional benefit could be derived from combination with selective chemotherapy.³³

i.p. RIT with ²¹²Pb has also been shown to be effective in treating vulvar squamous carcinoma. i.p. A-431 xenografts in athymic nude mice were treated with 370, 740, and 1,480 kBq of ²¹²Pb-35A7, an antibody that targeted non-internalizing anti-CEA receptors, or ²¹²Pb-trastuzumab that targeted ErbB2 (HER2) (internalizing) or PX (nonspecific control mAb), all given i.p. Transient hematological toxicities, as shown by leukocyte and platelet nadirs, were reported in mice treated with 1,480 kBq (70 MBq/kg) of ²¹²Pb-labeled mAbs. The median survival was higher (94 days) in mice treated with this dose of ²¹²Pb-35A7 than in those treated with ²¹²Pb-PX (18 days) and not reached (>130 days) in mice treated with 1,480 kBq ²¹²Pb-trastuzumab. These data suggested that non-internalizing anti-CEA and internalizing anti-HER2 could deliver ²¹²Pb to treat small-volume tumors, but toxicities were unavoidable using IgG-based RIT even at low doses, and internalization might offer therapeutic advantages.³⁴

In another study by Li et al. using astatine alpha-RIT, ²¹¹At-trastuzumab (anti-ErbB2 or anti-HER2) was examined in a model of HER2-positive peritoneal gastric cancer.³⁵ Biodistribution in N87 tumor-bearing mice showed that i.p. administration was more efficient than i.v., delivering 60% versus 18% injected dose per gram (ID/g), respectively to the tumor. A single 1 MBq i.p. dose was sufficient to eradicate disease in a third of treated mice, while reducing tumor burden in the rest. No significant toxicities were noted. Overall, the i.p. administration of ²¹¹At-trastuzumab significantly prolonged survival compared to trastuzumab alone.

Despite the significant progress in testing alpha-RIT in PC, dose-limiting toxicity remains an unsurmountable hurdle using standard forms of IgG for radioisotope delivery.

Preclinical pretargeted RIT in PC

The major limitation with RIT is the poor therapeutic index for many critical tissues (TI, ratio of radioactivity area under the curve (AUC) of tumor versus that of normal tissue) due to prolonged circulation and slow clearance of non-tumor-bound antibodies carrying toxic payloads.³⁶ To overcome this major hurdle, pretargeted RIT (PRIT) seeks to “pretarget” the tumor with a special antibody form that carries no radiohapten, which will be given only after unbound antibody is cleared from the blood, either by waiting or by using a clearing agent (Figure 2).³⁷ Several approaches to PRIT have been tested against PC over the decades (Table 1).

Figure 2.

Serial PET/CT imaging of a patient with DSRCT with intraperitoneal ¹²⁴I-omburtamab

Eligible patients diagnosed with DSRCT with peritoneal involvement were enrolled in a phase 1 trial of intraperitoneal radioimmunotherapy with ¹³¹I-omburtamab. After thyroid blockade and before radioimmunotherapy, patients received approximately 74 MBq of ¹²⁴I-omburtamab intraperitoneally. Shown above is the clearance of ¹²⁴I-omburtamab over a 144 h time period after injection. The above images show clearly the ability of PET/CT with intraperitoneally administered ¹²⁴I-omburtamab to image intraperitoneal DSRCT and eliminate well from the body.³⁸ Figure used with permission from Milan et al.³⁹

Beta-PRIT of PC

Beta-PRIT has similar benefits to beta-RIT, the largest difference of course being the method of targeting. This modality has been tested in various preclinical models with promise.

In a murine model of colorectal PC, nude mice were implanted i.p. with GPA33-expressing SW1222 cells to create an aggressive PC model. GPA33-PRIT was administered i.p. with an anti-GPA33/anti-DOTA BsAb, followed by an i.v. clearing agent, followed 4 h later by i.p. ¹⁷⁷Lu-S-2-(4-aminobenzyl)-DOTA. The efficacy and toxicity of single- versus three-cycle therapy (111 MBq/cycle) were evaluated in comparison to no-treatment control. Single-cycle treatment significantly prolonged median survival in comparison to controls, and the three-cycle therapy group survived beyond the study endpoint with at least one mouse, of 8, achieving histological cure at necropsy. There was no chronic myelosuppression or renal toxicity.⁴⁰ Maximal tolerated doses have yet to be defined.

Utilizing a different tumor target, the utility of a novel targeting molecule composed of four CC49-anti-tumor-associated glycoprotein 72 single-chain antibodies linked to streptavidin as a fusion protein (CC49 fusion protein) for treatment of i.p. LS174T CRC was studied. Mice were treated with CC49 fusion protein i.p. followed by a synthetic dextran clearing agent i.v., followed hours later by radiohapten i.v. ¹⁷⁷Lu-DOTA-biotin or ⁹⁰Y-DOTA-biotin. All doses of ¹⁷⁷Lu or ⁹⁰Y (400–600 μCi) significantly prolonged survival compared to no-treatment controls, but was not curative. No long-term toxicity was noted.⁴¹ Acute and chronic toxicities at curative doses have yet to be defined. More importantly, unintended renal uptake and immunogenicity of the streptavidin fusion protein emerged as major drawbacks of this PRIT approach.

PRIT based on biorthogonal click chemistry was investigated for the first time in the context of PC using CEA-targeting 35AZ mAb bearing trans-cyclooctene (TCO) moieties and several ¹⁷⁷Lu-labeled tetrazine (Tz) radioligands. Bioorthogonal click pretargeting is an elegant PRIT strategy because the chemical reaction between radiotherapeutic and targeting mAb occurs virtually instantaneously and is unaffected by the biological milieu. In a PC model of A431-CEA-Luc cells in nude mice, therapy with [¹⁷⁷Lu]Lu-Tz-2 was studied. It was found that an i.p. injection of 35A7-TCO + 40 MBq of [¹⁷⁷Lu]Lu-Tz-2 i.p. significantly slowed tumor growth compared to no-treatment control mice or those treated with [¹⁷⁷Lu]Lu-Tz-2 alone. Minimal toxicity was seen.⁴² Acute and chronic toxicities at curative doses require further study.

In a similar pretargeted study, a bispecific anti-CEA (CEACAM5; CD66e) × antihapten monoclonal antibody (TF2) was used in combination with a small peptide (IMP288) labeled with ¹⁷⁷Lu to treat a disseminated i.p. model of human CRC LS174T. Mice received TF2 i.v. and ¹⁷⁷Lu-IMP288 (60 MBq) i.v., non-pretargeted ¹⁷⁷Lu-IMP288, or saline. Images taken throughout therapy showed delayed growth in treated tumors, associated with prolonged survival, though no cure was noted.⁴³ Acute and chronic toxicities at curative doses have yet to be defined.

Beta-PRIT offers a new treatment model for PC, again with potential application to many carcinoma subtypes and antigen targets. There is early promise of prolonged survival with minimal toxicity.

Alpha-PRIT of PC

Though, in comparison to beta-PRIT, fewer studies have been done with alpha-PRIT, the preclinical models have shown great promise.

The concept of PRIT using alpha-emitting isotopes (alpha-PRIT) was tested in an identical model to that of beta-PRIT discussed earlier. Nude mice were implanted i.p. with GPA33-expressing SW1222 cells. Mice were treated with i.p. anti-GPA33/anti-DOTA BsAb followed by i.v. clearing agent, and later by single- versus two-cycle therapy with i.p. ²²⁵Ac-PrDOTA. In high-volume disease, prolonged survival was seen in the two-cycle regimen but not for single-cycle therapy, when compared to no-treatment controls. In low-volume disease, prolonged survival was seen in both single- and two-cycle therapy, with no significant difference between the two. Toxicity measurements found mild neutropenia at 1 month, which resolved.⁴⁴

Three-step PRIT has been simplified to 2-step PRIT that does not need clearing agent, utilizing a dual single-chain variable fragment (scFv) self-assembling dis-assembling (SADA) BsAb. This approach demonstrated specific tumor localization and reduced immunogenicity compared to IgG BsAb in mice because of the rapid renal clearance of the SADA. Most importantly, SADA bypasses normal gut uptake, a hurdle insurmountable so far clinically with IgG-based RIT targeting GPA33. Early work in CRC models has shown promise with high therapeutic index (TI) for both alpha- and beta-emitting radioisotopes.⁴⁵ No acute and chronic toxicities at curative tumor doses of >10,000 cGy were seen.

Most recently, 3-step PRIT was successfully simplified into a 2-step SADA approach⁴⁵ for i.p. CRC and AC. Two weekly injections of SADA (Step#1) and 74 kBq of [²²⁵Ac]Ac carried on Proteus (a derivative of DOTA-Bn)⁴⁶ achieved complete tumor responses. In CRC, median survival was significantly prolonged, 134+ days in treatment group versus 39–40 days among control groups; in AC, 6/10 mice in treatment group remained tumor free at the study endpoint of 180 days. Minimal myeloid, renal, hepatic, or neural toxicities were seen.⁴⁷

Though RIT has been successfully applied to ovarian cancer, fewer studies have been performed to study the effects of PRIT in ovarian cancer. In an i.p. ovarian xenograft model, 0.9 MBq ²¹¹At-MX35 i.p. RIT was compared with PRIT using avidin-MX35 i.p. in the first step followed i.p. by 1.0 or 1.5 MBq ²¹¹At-labeled, biotinylated, succinylated polylysine (²¹¹At-B-PLsuc). There was no difference in survival between PRIT and RIT; however, the 1.5 MBq PRIT group showed less ascites and smaller tumors 8 weeks after therapy than the RIT therapy group.⁴⁸ Acute and chronic toxicities at curative doses have yet to be defined.

Even though alpha-PRIT is a relatively new approach in the last decade, the toxicity and survival data so far in preclinical models spoke to its high clinical potential.

Theranostic applications in PC

Theranostics are agents that deliver radioactive material for the purpose of imaging/dosimetry besides therapy. PC in particular is known for its difficulty to be detected or adequately staged by PET or CT scan. Much of the RIT and PRIT research have potential utility as theranostics.

In a murine model of colorectal PC, nude mice were implanted i.p. with GPA33-expressing CRC SW1222. GPA33-PRIT was administered i.p. with a high-affinity anti-GPA33/anti-DOTA BsAb, followed by a clearing agent i.v., then 4 h later with i.p. ⁸⁶Y-DOTA-Bn payload for PET imaging. ⁸⁶Y is a positron-emitting isotope, with 33% positron abundance and half-life of 14.7 h, ideal for detecting both bulk tumor, as well as small tumor nodules in murine models of PC.⁴⁰ While ⁸⁶Y is suitable for tumor detection and dosimetry using PET, ⁹⁰Y is its therapeutic counterpart for beta-RIT or PRIT. This approach could have utility in the clinic for staging and treatment monitoring.

In a similar pretargeted study, a bispecific anti-CEA (CEACAM5; CD66e) × antihapten monoclonal antibody (TF2) i.v. was used in combination with a small peptide (IMP288) i.v. labeled with ¹¹¹In for gamma imaging, along with treatment with ¹⁷⁷Lu for beta-RIT, two isotopes as a theranostic pair. Single photon emission computed tomography (SPECT) imaging detected tumor uptake as early as 1 h post ¹¹¹In-IMP288 injection and correlated well with the uptake measured in dissected tumors.⁴³

In a study investigating PRIT based on biorthogonal click chemistry, again successful tumor visualization was established. In a PC model of A431-CEA-Luc tumor cells xenografted in nude mice, using i.p. CEA-targeting 35AZ mAB bearing TCO moieties + [¹⁷⁷Lu]Lu-Tz-2 i.p. achieved successful visualization by SPECT at 24 h post injection.⁴²

In mice bearing i.p. MISRII-positive tumors, i.p. 16F12 labeled with ⁸⁹Zr was able to localize to tumor with minimal background allowing accurate diagnosis with PET/CT at 24, 48, and 96 h post injection.³⁰

These experiments highlight the potential of radioligands not only in tumor therapy but also as methods to better identify disease burden in PC.

RIT of PC: Clinical studies

Despite the promise of RIT, its success and acceptance in clinical practice have been limited. While preclinical studies keep pointing to the insufficient TI (tumor-to-normal tissue absorbed dose ratios) for most RIT platforms, clinical testing continues to face similar headwinds. While tumor vs. normal tissue ratio at specific time points can serve well in tumor imaging, integrated activity ratios (TI or AUC) have to fulfill much more stringent criteria for therapeutic success because of organ-specific dose-limiting toxicities. Furthermore, TI has to be adequate not just for most organs but for all essential organs.⁴⁹ With improvement of targeting methods, mAb engineering, chemistry, and radioligands, the therapeutic index has seen steady improvement, and new clinical trials are slowly entering the clinic.

In a standard 3 + 3 phase 1 design for dose escalation, ²¹²Pb-TCMC-trastuzumab was delivered i.p. less than 4 h after administration of a normal tissue blocking i.v. dose of trastuzumab (4 mg/kg i.v.) in patients with PC who had failed standard therapies. Five dosage levels (7.4, 9.6, 12.6, 16.3, and 21.1 MBq/m²) showed minimal toxicity after follow-up of more than 1 year for the first group and more than 4 months for the others. The lack of substantial toxicity was consistent with the dosimetry assessments (mean equivalent dose to marrow, 0.18 mSv/MBq). Radiation dosimetry assessment was performed using pharmacokinetics data obtained in the initial cohort (n = 3). Limited redistribution of radioactivity out of the peritoneal cavity to circulating blood was seen, which cleared via urinary excretion, and no specific uptake in major organs was observed in 24 h. Tumor marker studies in patients with ovarian cancer showed a trend of decreasing tumor growth, accompanied by decreasing cancer antigen 72-4, aka tumor-associated glycoprotein 72, with increasing administered radioactivity. Other tumor markers, including carbohydrate antigen (CA125), human epididymis protein 4, serum amyloid A, mesothelin, interleukin-6, and CEA, did not correlate with imaging outcome.^50,51 However, there were no objective radiographic responses.

In another trial aimed at PC from ovarian cancer, ²¹¹At-MX35 F(ab’)2 was administered i.p. to patients in clinical remission after salvage chemotherapy. ²¹¹At, a relatively short-lived 7 h half-life alpha emitter, was given as an i.p. injection to facilitate rapid tumor targeting. Mean absorbed doses to normal tissues were calculated from clinical data, including blood and i.p. fluid samples, urine, γ-camera images, and single-photon emission CT/CT images. Extrapolation of preclinical biodistribution data combined with clinical blood activity data was used to estimate absorbed doses in additional tissues; the equivalent dose was calculated using an relative biological effectiveness (RBE) of 5.⁵² Using 100 MBq/L, organ equivalent doses were less than 10% of the estimated tolerance dose; urinary bladder, thyroid, and kidneys (1.9, 1.8, and 1.7 mGy per MBq/L) received the highest bystander doses. Safety concerns warrant further optimization.^49,52 In a dose-escalation study with long-term follow-up (>10 years), the activity concentration was escalated to 1.7 L of 215 MBq/L (5 MBq/kg, assuming a standard 70 kg patient) without any dose-limiting toxicities and no observed hematologic toxicity. Median survival was 35 months. Calculations of the absorbed doses showed that a lower specific activity is associated with a lower single-cell dose, whereas a high specific activity may result in a lower central dose in microtumors.⁵³

In a phase 1 study treating desmoplastic round cell tumor (DSRCT), a rare sarcoma of adolescents/young adults primarily involving the peritoneum, ¹³¹I-omburtamab (¹³¹I-8H9) directed at B7-H3 was investigated. After thyroid blockade, 52 patients received ¹³¹I-omburtamab i.p. at escalated activities from 1.11 to 3.33/GBq/m² based on PET dosimetry using 74 MBq ¹²⁴I-omburtamab i.p. in each patient. As shown in Figure 2, PET/CT imaging with i.p. administered ¹²⁴I-omburtamab enables the assessment of i.p. distribution and estimation of I-131 absorbed dose to peritoneal space and normal organs before therapy. MTD was not reached; there were no dose-limiting toxicities. Mean projected absorbed normal organ doses for ¹³¹I-omburtamab based on PET/CT were low and well within tolerable limits.^38,39

To date, there have been few clinical trials with RIT in PC. These early-stage trials show radioactivity uptake in tumor tissue with tolerable normal organ doses. It is notable that besides the many targets and tumor types that have yet to be explored, novel antibody platforms continue to be developed to widen the therapeutic window, providing opportunities for more effective clinical application in the coming decades.

Conclusion and future directions

A wealth of data have been gathered in preclinical i.p. murine models, summarized in Table 1, that highlight the promise of both RIT and PRIT, against a variety of cancers including gastric, colorectal, ovarian, and pancreatic, targeting different antigens using diverse radionuclides. Although minimal hematological toxicity was observed in most models except for transient leukocytosis and weight loss, significant toxicities were common at doses in order to achieve durable remissions.

A few early clinical trials have been reported (Table 2) that showed PC RIT was safe at sub-therapeutic doses, but they were unable to dose escalate because of concerns about normal organ toxicity so that suboptimal or nondurable tumor responses were seen. In addition, many patients were treated at the time of minimal residual disease (or remission) making it hard to evaluate efficacy in the absence of a randomized arm. Another reason for the scarcity of randomized or single-arm clinical trials of RIT or PRIT is the competition for patients with PC in established oncology practice. Besides the necessity to relearn or retrain, oncology personnel or oncology centers must be equipped to deal with radioisotopes and containment of radiohazards. One important consideration of PRIT is the ability to introduce a shared workflow in which oncologists could administer the non-radioactive antibody while the nuclear medicine specialists administer the radioisotopes, thereby encouraging a multidisciplinary effort. Future studies need to pay attention to these shortcomings if such therapeutics are to be widely accepted by the medical community.

Table 2.

Summary of clinical trials of RPT for peritoneal metastases

A summary of clinical trials over the past decade of RIT or PRIT in PC of various cancer types.

Normal organ toxicities from unintended damage to non-tumor-bearing tissue are a concern for many treating physicians. One such preclinical study looked at the toxicity of i.v. versus i.p. injections with “free” ²¹²Pb, the purpose being to study how the isotope would behave should it become unbound from its chelator. While i.v. injections cause moderate hepatocyte toxicity, i.p. injections have slower hepatic exposure, subsequently causing less toxicity.⁵⁴ However, such pharmacokinetics of free radiometals may not be generalizable. The use of single bolus of cold-mAb (with no radioisotope) before RIT has been used to enhance tumor uptake and dose distribution⁵⁵; however, its utility varies among antibodies and targets since the tumor %ID/g could be compromised. Another concern is the possibility of late effects in treated patients, especially after exposure to powerful genotoxic radionuclides, including alpha emitters, which could cause second cancers especially among patients carrying cancer predispositions somatically or in their germline. One epidemiologic study on the lifelong detriments after alpha-particle irradiation tried to link dosimetry to the risk of second cancer after ²¹¹At-based i.p. RIT—a risk estimated at 1.53 excess cases of second cancer per 100 patients treated. In this study, 10 organs were at risk, primarily the urinary system. Since most isotopes are eliminated through the urine, measures to reduce dose to the urinary tract should decrease second cancer risk.⁵⁶

In summary, PC is a lethal complication observed during the clinical course of many common GI malignancies, and effective therapy remains a major unmet need. The list of candidate RPT is growing, some showing promise in laboratory models of disseminated peritoneal cancers both for improving diagnosis/staging as well as palliation and some even cures in select settings. Also, at least in human xenograft studies in rodents, relative safety of the RPT using alpha particles could exploit the precision and potency of high LET particle therapy to complement standard chemotherapy treatments without pushing the toxicity limits. We suggest that PRIT methodologies have advanced to the point where a harmonized approach can be more widely applied, taking into account pharmacokinetic considerations to overcome previous PRIT failures. Beyond the supply chain issues of individual radioisotopes, the challenge continues, i.e., (1) to widen the therapeutic window (MTD vs. MED), not just TI of individual organs, to reduce both acute and late effects, (2) to design trials with harmonized endpoints, both preclinical and clinical, bringing together a team of surgical oncologist, medical oncologist, nuclear medicine physicians, chemists and radiochemists, medical physicists, and pharmacologists, and (3) to achieve PC cures beyond just GI cancers.

Acknowledgments

N.-K.V.C. was supported partly by Enid A. Haupt Endowed Chair, the Robert Steel Foundation, and the NCI Cancer Center Support Grant P30 CA008748.

Declaration of interests

Both Memorial Sloan Kettering (MSK) and N.-K.V.C have financial interest in Y-mAbs and Eureka Therapeutics. N.-K.V.C. reports receiving past commercial research grants from Y-mAbs Therapeutics. N.-K.V.C was named as the inventor on multiple patents filed by MSK, including those licensed to Y-mabs Therapeutics and Biotec Pharmacon.

S.M.L. reports receiving commercial research grants from Y-mAbs Therapeutics, Inc.; Genentech, Inc.; WILEX AG; Telix Pharmaceuticals Limited; and Regeneron Pharmaceuticals, Inc.; holding ownership interest/equity in Elucida Oncology, Inc., and holding stock in ImaginAb, Inc., and Y-mAbs Therapeutics. S.M.L. is the inventor of issued patents both currently unlicensed and licensed by MSK to Samus Therapeutics, Inc.; Elucida Oncology, Inc.; and Y-mAbs Therapeutics, Inc. S.M.L. serves or has served as a consultant both compensated and uncompensated to Cynvec, LLC; Eli Lilly &Co.; Prescient Therapeutics Limited; Advanced Innovative Partners, LLC; Gerson Lehrman Group; Progenics Pharmaceuticals, Inc.; Exini, Inc.; and Janssen Pharmaceuticals. S.M.L. is a scientific consultant to Medimagemetric LLC and co-inventor of provisional patent for Soothsayer: number 63/193,700.

S.M.C. serves as a consultant to Affibody AB.

N.-K.V.C., S.M.L., D.R.V., and S.M.C. were named as inventors in the following patent applications relating to GPA33: SK2014-074, SK2015-091, SK2017-079, SK2018-045, SK2014-116, SK2016-052, and SK2018-068 filed by MSK. S.M.L., N.-K.V.C., D.R.V., and S.M.C. were named as inventors in PCT/US2021/039418 (THOR cell [tumor homing radio-emitting cell]).