[¹⁷⁷Lu]Lu-FAPI-2286 radioligand therapy in heavily pre-treated patients with advanced breast and gastrointestinal cancers: a single-center retrospective experience

Abstract

Background

Breast and gastrointestinal (GI) cancers remain leading causes of cancer-related mortality worldwide, particularly in the metastatic setting, where therapeutic options are limited, and drug resistance inevitably develops. This study aims to evaluate the feasibility and safety of [¹⁷⁷Lu]Lu-FAPI-2286 in patients with breast and GI cancers.

Methods

In this retrospective, single-center, observational analysis, we evaluated the safety, tolerability, and preliminary efficacy of [¹⁷⁷Lu]Lu-FAPI-2286 in 14 patients with advanced metastatic breast (n = 10) and GI (n = 4) malignancies who had exhausted standard therapies. Adverse events were graded per CTCAE. Clinical response was assessed by evaluating symptomatic outcomes and biochemical markers, and radiographic assessment was performed using post-therapy imaging.

Results

The cohort (median age 46.5 years) was heavily pre-treated, with extensive bone, liver, and lung metastases. [¹⁷⁷Lu]Lu-FAPI-2286 was generally well tolerated; hematologic toxicity included Grade 3 anemia in two patients and one case of Grade 4 thrombocytopenia. Two patients experienced transient post-administration pain flares. Symptomatic pain relief was reported in four patients (28.6%), particularly among those with predominant bone metastases. Onset occurred approximately one week after therapy and lasted up to one month between cycles. Radiographic outcomes showed stable disease (SD) in 28.6% and progressive disease (PD) in 21.4% of evaluable patients; however, 50% of patients could not be evaluated radiographically due to rapid clinical decline. Median follow-up was 3.25 months (range: 2-17.5 months).

Conclusions

In this study, [¹⁷⁷Lu]Lu-FAPI-2286 demonstrated feasibility and an acceptable safety profile, with meaningful palliative benefits in a subset of patients. No objective responses were observed, and the primary benefit was palliative.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12885-026-16270-1.

Introduction

Breast and gastrointestinal (GI) malignancies constitute a significant share of the global cancer burden. Breast cancer remains the most frequently diagnosed cancer in women worldwide and a leading cause of cancer-related mortality. Similarly, colorectal and gastric cancers are among the leading contributors to global cancer deaths, marked by high rates of metastatic spread and recurrence [1]. While early-stage disease is potentially curable, managing metastatic disease remains a formidable clinical challenge. Standard-of-care protocols, including surgery, cytotoxic chemotherapy, radiotherapy, and targeted agents, have significantly improved survival outcomes. However, the development of therapeutic resistance is almost inevitable [2, 3].

Once patients progress on standard lines of therapy, the therapeutic landscape narrows. Patients with refractory disease, those who have exhausted approved chemotherapeutic, immunotherapeutic, and targeted options, face a dismal prognosis and limited palliative alternatives. In this end-stage setting, tumor heterogeneity often renders mutation-specific treatments ineffective, creating an urgent unmet need for novel therapeutic targets that are pan-tumoral and independent of specific driver mutations. This clinical void has catalyzed a shift toward targeting the tumor microenvironment (TME) rather than the malignant cells alone [4, 5].

A key component of the TME in epithelial malignancies, particularly in breast and GI cancers, is the presence of Cancer-Associated Fibroblasts (CAFs). These stromal cells contribute to tumor growth, immune evasion, and structural desmoplasia, often acting as a physical barrier to drug delivery. Fibroblast Activation Protein (FAP) is a type II transmembrane serine protease that is highly overexpressed on the cell surface of CAFs in over 90% of epithelial carcinomas but is virtually absent in healthy adult tissues [6–8]. This differential expression makes FAP an ideal target for theranostic approaches. Diagnostic imaging with Gallium-68-labeled FAP inhibitors has demonstrated superior tumor-to-background contrast compared with standard [¹⁸F]FDG PET/CT in several indications, including lobular breast cancer, signet-ring cell gastric carcinoma, and peritoneal carcinomatosis, confirming the high abundance of this target in these difficult-to-treat entities [9, 10].

The intense tracer uptake observed on diagnostic Fibroblast Activation Protein Inhibitor (FAPI) PET/CT provided the rationale for FAP-targeted Radioligand Therapy (RLT) with beta-emitting isotopes such as Lutetium-177. The therapeutic mechanism relies on the cross-fire effect, in which radiation emitted by the nuclide bound to stromal CAFs penetrates and destroys adjacent tumor cells, bypassing the need for tumor cells to express the target [5, 11].

While early studies used monomeric peptides such as FAPI-04 and FAPI-46, recent work has focused on FAPI-2286, a novel cyclic peptide. Preclinical and early clinical data suggest that FAPI-2286’s cyclic structure confers higher binding affinity and, crucially, prolonged tumor retention compared with its monomeric predecessors. This enhanced pharmacokinetic profile theoretically maximizes the absorbed radiation dose to the tumor while maintaining a manageable safety profile, making it a promising candidate for systemic radionuclide therapy [12].

Cyclic FAPI-2286 is expected to outperform monomeric FAPI tracers because its cyclic peptide structure confers higher binding affinity, structural rigidity, and resistance to enzymatic degradation, resulting in more stable target interactions [13]. In addition, this architecture yields prolonged intratumoral retention and sustained tumor uptake, which are critical for improving therapeutic efficacy compared with rapidly clearing monomeric agents [14].

Despite theoretical advantages, clinical data on the safety and efficacy of [¹⁷⁷Lu]Lu-FAPI-2286, particularly in a salvage setting for diverse solid tumors, remain limited. This retrospective study aims to evaluate the safety, tolerability, and preliminary therapeutic efficacy of [¹⁷⁷Lu]Lu-FAPI-2286 in a cohort of heavily pre-treated patients with advanced breast and gastrointestinal malignancies who had no remaining standard therapeutic options. We focus on clinical outcomes, including pain palliation and discordance between stromal and epithelial response markers.

Materials and methods

Study design and ethical considerations

This retrospective, single-center, observational study evaluated the feasibility, safety, and efficacy of fibroblast activation protein inhibitor (FAPI) radioligand therapy (RLT) with [¹⁷⁷Lu]Lu-FAPI-2286. The study included patients treated between 2024 and 2026 on a compassionate-use basis and referred to our department. Of the cohort, 2 breast cancer patients with a favorable palliative effect on bone pain have been previously reported [15].

The protocol was conducted in accordance with the ethical principles of the Declaration of Helsinki. All patients provided written informed consent to receive the unapproved radiopharmaceutical, acknowledging the experimental nature of the therapy, potential risks, and the collection of their clinical data for research purposes. Patients were carefully selected by multidisciplinary teams.

Patient population

A total of 14 patients with histologically confirmed advanced malignancies were included. Eligibility criteria for [¹⁷⁷Lu]Lu-FAPI-2286 therapy were as follows: Exhaustion of Standard Therapies: Patients must have progressed on or been ineligible for standard-of-care treatments, including surgery, chemotherapy, radiotherapy, and immunotherapy. FAPI-Avidity: positive tumor uptake on diagnostic [^99mTc]Tc-FAPI SPECT/CT (tumor-to-background ratio > 1), defined as uptake intensity exceeding background levels. Performance Status: Eastern Cooperative Oncology Group (ECOG) performance status of 0–3 (all patients had an initial ECOG of 0). Organ Function: adequate baseline hematological, renal, and hepatic function, unless organ dysfunction was directly attributable to tumor involvement and the potential benefit outweighed the risk.

A notable methodological consideration in this study is the use of [⁹⁹ᵐTc]Tc-FAPI-46 SPECT/CT rather than [⁶⁸Ga]Ga-FAPI PET/CT for patient selection, as recommended by the SNMMI/EANM guideline [9]. At our institution, access to ⁶⁸Ga generators and ⁶⁸Ga‑labeled FAPI tracers remains limited by regulatory constraints, high production costs, and the logistical complexity of radiopharmaceutical synthesis. SPECT/CT inherently offers lower spatial resolution and sensitivity than PET, which may result in under‑detection of small or weakly FAP‑avid lesions and the potential inclusion of patients with borderline tumor uptake. Conversely, using a tumor‑to‑background ratio exceeding 1 as the positivity criterion, as was done here, likely enriches the cohort for highly FAP‑expressing tumors, making it clinically relevant for therapeutic decision‑making despite reduced sensitivity.

Radiopharmaceutical preparation and administration

The radiolabeling of [¹⁷⁷Lu]Lu-FAPI-2286 was performed at Pars Isotope Company in Tehran, Iran, using the procedure outlined below. A solution containing 210 mg of sodium L-ascorbate, 42 mg of gentisic acid, and 170 µg of FAP-2286 in 1 mL of 0.04 M HCl (pH 4.5) was combined with 7400 MBq of ¹⁷⁷LuCl3 in 0.5 mL of 0.05 M HCl. The mixture was heated to 95 °C for 60 min, then diluted with 0.9% saline and filtered under sterile conditions. Subsequently, tests were conducted to assess radiochemical purity, pH, endotoxin levels, and sterility. The [¹⁷⁷Lu]Lu-FAPI-2286 produced in this study exhibited excellent characteristics, including radiochemical purity greater than 99%, radionuclide purity exceeding 99.9%, endotoxin levels below 17.5 EU/mL, and a pH between 4 and 5.

Upon receipt in our nuclear medicine department, the total activity of the radiopharmaceutical was measured with a dose calibrator, and the patient-specific dose was dispensed before administration. The treatment was administered intravenously as a slow bolus over 10–15 min. The administered activity was determined empirically based on patient weight (adjusted for patients weighing less than 50 kg) and bone marrow reserve (adjusted for Hb of 8–9 g/dL and Plt of 100–120/mm³). The administered activity per cycle ranged from 3.7 GBq (100 mCi) to 7.4 GBq (200 mCi). Cycles were repeated at intervals of approximately 4 to 8 weeks, depending on hematological recovery and clinical status. Unlike DOTATATE-based therapies, amino acid co-infusion for renal protection was not routinely mandated due to the low physiological renal retention of FAPI-2286.

Assessments and data collection

Patients were monitored for adverse events (AEs) during and immediately after administration, including pain flare and allergic reactions. Laboratory assessments were performed at baseline and at follow-up visits. Key parameters monitored included: Hematology: Hemoglobin (Hb), White Blood Cell (WBC) count, and Platelets (Plt); Serum Creatinine (Cr); Aspartate Transaminase (AST), Alanine Transaminase (ALT), Alkaline Phosphatase (ALP), and Lactate Dehydrogenase (LDH). Toxicity was graded according to the National Cancer Institute Common Terminology Criteria for Adverse Events (NCI-CTCAE) version 5.0.

Clinical and biochemical response

Clinical response

Symptomatic changes were assessed using patient-reported outcomes, with a focus on cancer-related pain. Pain relief was qualitatively categorized (Considerable relief, Stable, or Worsening) and recorded for onset and duration.

Biochemical response

Serum tumor markers relevant to the primary histology were measured serially. A biochemical response was defined as a decrease in marker levels, whereas progression was defined as a sustained increase.

Radiographic response assessment

In this study, treatment response evaluation relied on the radio-oncologist’s and nuclear physician’s judgments for each patient on a case-by-case basis, based on signs and symptoms, tumor markers (if available), and post-treatment SPECT/CT. Conventional imaging (CT scan, ultrasonography) was used as needed, and the ^99mTc-FAPI-46 scan was typically performed after the 4th cycle.

Post-therapy imaging (whole-body scintigraphy or SPECT/CT) was performed approximately 24 h after administration to verify radioligand biodistribution and tumor retention. For disease monitoring, restaging was performed using RECIST 1.1 criteria and, where feasible, by assessing post-therapy scintigraphy lesion uptake with Krenning-like scoring.

Tumor response on ^99mTc-FAPI-46 SPECT/CT was classified qualitatively as: Complete Response (CR), defined as the disappearance of all target lesions; Partial Response (PR), defined as a significant reduction in tumor burden; Stable Disease (SD), defined as neither sufficient shrinkage to qualify for PR nor sufficient increase to qualify for PD; Progressive Disease (PD), defined as the appearance of new lesions or unequivocal progression of existing lesions; and unverifiable: patients who died due to disease or clinically deteriorated before follow-up imaging could be performed.

Statistical analysis

Because the study design was retrospective, all eligible patients were included in the analysis, and no formal sample size calculation was performed. Descriptive statistics were used to summarize the data. Continuous variables are presented as mean ± standard deviation, and categorical variables as frequencies and percentages. Overall survival (OS) was estimated using the Kaplan–Meier method and defined as the interval from the first administration of [¹⁷⁷Lu]Lu-FAPI-2286 to the date of death or last follow-up. Survival curves were generated to assess patient outcomes over the follow-up period. Statistical analyses were performed using SPSS version 27.0, and a p-value < 0.05 was considered statistically significant.

Results

Patient characteristics

A total of 14 patients with histologically confirmed advanced and metastatic malignancies were enrolled in this study. The cohort comprised 11 females (78.6%) and 3 males (21.4%), with a median age of 46.5 years (range: 31–65 years). The predominant primary malignancy was breast cancer (n = 10, 71.4%), encompassing molecular subtypes including Luminal A, Luminal B, and Triple-Negative Breast Cancer (TNBC). The remaining cohort (n = 4, 28.6%) presented with gastrointestinal malignancies, specifically colorectal adenocarcinoma (n = 2), gastric poorly differentiated adenocarcinoma (n = 1), and general gastrointestinal adenocarcinoma (n = 1).

The study population was heavily pre-treated; all patients had received prior standard-of-care therapies, including surgery and chemotherapy. Radiotherapy had been administered to 11 patients (78.6%), and immunotherapy was used in a minority of cases. At baseline, distant metastases were extensive, with bone (n = 9), liver (n = 4), and lung (n = 5) the most common sites of metastatic involvement. A detailed summary of baseline demographics and clinical characteristics is presented in Table 1 and Supplementary Table 1.

Table 1.

Baseline Characteristics of Patients Treated with [¹⁷⁷Lu]Lu-FAPI-2286

Patient ID	Sex	Cancer	Subtype	Previous related Surgery	Previous Chemotherapy	Previous RT	Sessions	Concomitant treatment	Any patient reported Side effects	Pain Changes during Tx	Primary lesion
1	M	rectum	adenocarcinoma	yes	yes	yes	3	No	No	NR	NS
2	M	gastric	poorly differentiated adenocarcinoma	no	yes	no	3	no	No	NR	NS
3	F	colon	adenocarcinoma	no	yes	no	2	yes	no	NR	NS
4	M	colon	adenocarcinoma	yes	yes	no	1	yes	no	NR	NS
5	F	breast	NS	yes	yes	yes	3	no	thrombocytopenia	Considerable pain relief for 3–4 weeks after each cycle (improve pain from VAS score of 8 to 2)	resected
6	F	breast	NS	yes	yes	yes	1	no	no	Considerable pain relief up to 4 weeks after each cycle (improved pain from VAS score of 7 to 2)	resected
7	F	breast	IDC, luminal B	yes	yes	yes	1	no	infection/sepsis	NR	locally advanced mass
8	F	breast	IDC /luminal A	yes	yes	yes	1	no	increase pain in early days after treatment	Acute and transient increase pain in early days after treatment (increase in VAS score from 4 to 7 for about a week)	resected
9	F	breast	IDC, luminal B	yes	yes	yes	2	no	no	NR	recurrent mass
10	F	breast	IDC, luminal B	yes	yes	yes	1	no	no	NR	NS
11	F	breast	IDC/luminal B	yes	yes	yes	2	no	no	Considerable pain relief up 4 weeks after each cycle (improve in pain from VAS score of 7 to 2)	resected
12	F	breast	luminal B	no	yes	no	2	no	no	NR	one mass in the left breast
13	F	breast	IDC/triple -	yes	yes	yes	1	yes	Tx cessation due to pain increase in the primary site	Pain increase in primary tumor location, attributed to the disease progression (increase in VAS score from 3 to 7)	multiple FAPI-avid lesions in left breast extending to left lower thoracic region and one new lesion in right breast
14	F	breast	IDC, Grade III/triple negative	yes	yes	yes	1	no	no	NR	mass lesions, fat stranding and edema extending to right lower hemithorax and right arm due to tumoral involvement.

NR not reported, NS not specified, SD stable disease, PD progressive disease, OS overall survival, IDC Invasive Ductal Carcinoma, mets metastasis, LNM Lymph Node Metastasis

Treatment administration

Patients received FAPI radioligand therapy labeled with Lutetium-177 ([¹⁷⁷Lu]Lu-FAPI-2286). The administered activity per cycle ranged from 3.7 GBq (100 mCi) to 7.4 GBq (200 mCi). The administered dose was determined by the clinician’s judgment and the patient’s status, and was adjusted for body weight (less than 50 kg), creatinine level (adjusted by eGFR of less than 60–65 ml/min/1.73 m²), baseline blood counts (adjusted for Hb of 8–9 g/dl, absolute neutrophil count of 1-1.5 /mm³, and platelet count of 100–120 /mm³), and advanced age (more than 75 years). The median number of cycles administered was 1.5 (range: 1–3). Treatment continuation was determined by clinical status and disease progression; 3 patients (21.4%) completed 3 cycles, 5 patients (35.7%) completed 2 cycles, and 6 patients (42.9%) received a single cycle. One patient required treatment cessation after the first cycle due to a significant increase in pain at the primary tumor site.

Safety and tolerability

Overall, [¹⁷⁷Lu]Lu-FAPI-2286 was well tolerated, with a manageable safety profile. Hematological parameters, including hemoglobin (Hb), white blood cell (WBC) count, and platelet (Plt) count, were assessed at baseline and after each cycle.

As detailed in Table 2, mean hemoglobin levels remained relatively stable from baseline (11.43 ± 1.16 g/dL) to post-cycle 1 (11.35 ± 1.27 g/dL). However, Grade 2 anemia was observed in 3 patients. Of these, only one patient had a baseline Hb level below 10 and continued to have low levels. Thrombocytopenia was the most notable hematological adverse event. The cohort mean platelet count decreased slightly from 234 × 10³/µL to 203 × 10³/µL after the first cycle. One patient—a 31-year-old woman with breast cancer and no evidence of bone marrow involvement at baseline—developed severe cumulative thrombocytopenia. Her initial CBC was apparently normal, with a platelet count of 160 × 10³/µL. After the first cycle, her platelet count decreased to 125 × 10³/µL; it further declined to 35 × 10³/µL after the second cycle and reached a nadir of 15 × 10³/µL after the third cycle, meeting the criteria for Grade 4 toxicity.

Table 2.

Baseline and after 1 cycle laboratory results of Patients Treated with [¹⁷⁷Lu]Lu-FAPI-228

	Baseline				After Cycle 1				Wilcoxon signed rank test
	Level (n)	CTCAE Toxicity			Level (n)	CTCAE Toxicity			Paired number	Z	p-value
		G1	G2	G3		G1	G2	G3
Hb (g/dl)	11.43 ± 1.16 (13)	7	2	0	11.35 ± 1.27 (14)	7	2	0	12	0.196	0.875
WBC (/mm3)	6.5 ± 4.84 (12)	4	0	0	6.93 ± 3.98 (12)	3	0	0	11	-0.178	0.894
Plt (/mm3)	234.31 ± 104.13 (13)	2	0	0	203.77 ± 81.11 (13)	3	0	0	12	0.549	0.622
eGFR (ml/min/1.73 m2)	93.81 ± 23.75 (14)	3	0	1	96.34 ± 20.79 (13)	4	1	0	13	-0.445	0.689

Renal and hepatic functions were generally preserved. Using the Wilcoxon signed-rank test, there were no statistically significant elevations in serum creatinine, aspartate transaminase (AST) (from 43.86 ± 48.14 at baseline to 57.2 ± 66.8 after the first cycle, Z = -0.674, p = 0.475), alanine transaminase (ALT) (from 35 ± 35.39 at baseline to 40.56 ± 32.28 after the first cycle, Z = 0.000, p > 0.05), or alkaline phosphatase (ALP) (from 318.79 ± 200.05 at baseline to 471 ± 367.43 after the first cycle, Z = -1.606, p = 0.128) across the cohort (Table 2). However, transient clinical pain flares were reported. Two patients experienced an acute increase in pain intensity at tumor sites shortly after administration, which required clinical management or, in one case, led to treatment discontinuation (Fig. 1).

Fig. 1.

A middle-aged woman with a history of left breast cancer and extensive skeletal metastases was initially treated with systemic chemotherapy. Subsequent lines of therapy included Palbociclib followed by Exemestane. Given ongoing bone metastasis-related pain, the patient was referred for ¹⁷⁷Lu-FAPI-2286 therapy within a clinical trial. A baseline ^99mTc-FAPI-46 SPECT/CT scan performed before treatment demonstrated multiple FAPI-avid skeletal metastases involving the spine, ribs, sternum, humeri, scapulae, femora, and pelvic bones. Additionally, a FAPI-avid soft-tissue mass in the left breast was identified, consistent with the primary tumor (A). The patient subsequently underwent two cycles of ¹⁷⁷Lu-FAPI-2286 therapy. Following treatment, she experienced partial improvement in bone pain. However, the second post-therapy scan demonstrated disease progression, with an increased number and extent of skeletal metastases (B-C). Ultimately, due to diffuse bone marrow involvement, the patient developed pancytopenia, leading to discontinuation of ¹⁷⁷Lu-FAPI-2286 therapy

Clinical and symptomatic response

Despite the advanced stage of disease, significant symptomatic relief was reported in a subset of patients. Four patients (28.6%) reported distinct improvements in cancer-related pain. For instance, a rectal cancer patient reported reduced back pain and improved ambulation. Similarly, three breast cancer patients reported considerable to significant pain relief, with onset approximately one week post-therapy and lasting up to one month between cycles.

Biochemical response was assessed using serum tumor markers (CEA, CA19-9), where applicable. Discordance between biochemical markers and imaging findings was observed. For example, a patient with rectal cancer showed a rise in CEA from a baseline of 39.23 to 42 ng/mL after the 3rd cycle during treatment, suggesting molecular progression despite an initial qualitative assessment of SD (Fig. 2).

Fig. 2.

A middle-aged patient with a history of rectal cancer diagnosed six years earlier underwent surgical resection followed by chemoradiotherapy. Six months before the current presentation, he developed abdominopelvic pain. MRI and ¹⁸F-FDG PET/CT demonstrated local tumor recurrence with invasion of adjacent structures, sacral skeletal metastasis, and suspected pulmonary metastases. According to the multidisciplinary tumor board decision, ¹⁷⁷Lu-FAPI therapy was initiated in combination with chemotherapy as part of a clinical trial. Baseline ^99mTc-FAPI SPECT/CT showed intense FAPI uptake in the right seminal vesicle and the right posterolateral wall of the urinary bladder (Red arrow in B), with invasion of the right sacral ala (Yellow arrow in A, B). FAPI uptake was also noted along the orthopedic devices (Green arrow in B), and suspicious pulmonary metastases without FAPI uptake were present (White arrow in B). The patient received four cycles of ¹⁷⁷Lu-FAPI therapy (C-F). After the final cycle, disease progression was observed. This was characterized by newly developed FAPI-avid lesions in the lungs (White arrow in G) and mediastinum (Green arrow in H), involvement of the right scapula (Yellow arrow in I), and progression of the lytic lesion in the sacrum (Red arrow in J)

Radiographic response was assessed qualitatively using post-treatment SPECT/CT or restaging imaging: 4 patients (28.6%) achieved stable disease (SD), 3 patients (21.4%) had confirmed progressive disease (PD), and in 7 patients (50%), radiographic response could not be definitively classified due to rapid clinical deterioration, loss to follow-up, or lack of comparable cross-sectional imaging.

Survival outcomes

As of the data cutoff, 10 patients (71.42%) had died from their disease, including three with GI cancer and seven with breast cancer. Four patients (28.58%) were still alive. Follow-up time from the initial injection ranged from 2 to 17.5 months, with a median of 3.25 months. The median survival time for breast cancer patients and for all patients was 6 months, 95% CI: [0.00-13.57] and 6 months, 95% CI: [0.00-12.69], respectively; the mean OS for GI cancer patients was 4.63 months, 95% CI: [2.78–6.47] (Fig. 3).

Fig. 3.

Kaplan-Meier Curve of OS according to the tumor type

Discussion

The present study evaluated the safety, feasibility, and preliminary efficacy of [¹⁷⁷Lu]Lu-FAPI-2286 in a cohort of 14 patients with end-stage, heavily pre-treated metastatic malignancies, predominantly breast and gastrointestinal cancers. The results suggest that [¹⁷⁷Lu]Lu-FAPI-2286 is generally well-tolerated, with a manageable safety profile, but its efficacy as a salvage therapy in late-stage disease is heterogeneous. Our findings highlight significant symptomatic relief in a subset of patients despite the challenges of rapid disease progression, emphasizing the complex role of targeting the TME in advanced oncology.

FAP is highly expressed in CAFs, which constitute a major component of the tumor stroma but are largely absent in healthy adult tissues. Unlike other RLTs such as [¹⁷⁷Lu]Lu-PSMA or [¹⁷⁷Lu]Lu-DOTATATE that target tumor cells directly, FAPIs target the tumor’s scaffolding and life-support system [16]. In tumors that have been heavily pre-treated but have survived, resistant cell clones emerge, making future therapies more difficult and less effective [17, 18].

As patients’ bone pain subsides, the promising activity of FAPI for bone metastases, especially in breast cancer patients, can be supported, making it a promising targeted therapy for bone metastatic lesions. In this cohort, stable disease (SD) was achieved in 28.6% of patients, and significant pain relief was reported in 28.6% (n = 4), predominantly among those with extensive bone metastases. The mechanism of pain relief, even in the absence of objective radiographic regression, may be attributed to stromal modulation. By irradiating the CAFs, [¹⁷⁷Lu]Lu-FAPI-2286 may disrupt paracrine signaling loops, such as TGF-β and CXCL12, that drive inflammation and peritumoral edema, thereby reducing pressure on nociceptors in the periosteum or soft tissues. As previously hypothesized, Lu-FAPI does not directly target tumor cells; it only destroys the tumor’s castle, as supported by the SD state and pain relief observed in several cases. Thus, the presence of standard therapies, such as immunotherapy or chemotherapy regimens, which have been seen in only 3 of our cohort patients, is necessary to affect the unprotected, vulnerable tumor cells [16].

However, the high rates of unverifiable responses (50%) and confirmed progressive disease (21.4%) underscore the aggressive nature of the malignancies treated. Most patients had exhausted all standard lines of therapy, including chemotherapy, radiotherapy, and immunotherapy. The rapid clinical deterioration in half the cohort suggests that for FAP-targeted RLT to be effective, it may need to be administered earlier in the disease course, before the tumor burden exceeds the threshold at which stromal depletion can induce tumor collapse.

The radioligand used in this study, FAPI-2286, uses a cyclic peptide structure, which theoretically offers higher binding affinity and prolonged tumor retention compared to earlier-generation monomers such as FAPI-04 or FAPI-46 [12]. Although our study did not conduct head-to-head dosimetric comparisons, sustained pain relief lasting up to one month between cycles in responders supports the hypothesis of prolonged tumor residence time. Nevertheless, clinical benefits were transient for many, suggesting that the radiation absorbed dose delivered by the current activity (3.7–7.4 GBq) may still be insufficient to sterilize bulky, hypoxic tumors where CAFs are rapidly replenished.

The safety profile observed in this study aligns with previous FAPI-RLT reports, showing a generally low incidence of nonhematological grade 3/4 toxicities. Renal and hepatic functions were largely preserved, consistent with the low physiological background uptake of FAPIs in the liver and kidneys [19].

However, hematotoxicity remains a concern, particularly in patients with a high skeletal tumor burden. We observed a general trend of mild hemoglobin decline and thrombocytopenia. Notably, a 31-year-old breast cancer patient developed Grade 4 thrombocytopenia: her baseline full blood count was entirely normal, and there was no imaging or clinical evidence of bone marrow infiltration. Her platelets declined from 160 × 10³/µL at baseline to 125, 35, and 15 × 10³/µL after the first, second, and third cycles, respectively. This is likely due to the cross-fire effect of beta-particles from ¹⁷⁷Lu irradiating the healthy bone marrow adjacent to FAP-avid skeletal metastases.

This case demonstrates that, even without pre-existing marrow compromise, the cumulative myelotoxicity from repeated [¹⁷⁷Lu]Lu-FAPI-2286 administrations can be profound. The prolonged tumor retention time of the cyclic peptide, while therapeutically advantageous, may increase the cumulative bone marrow absorbed dose over successive cycles, particularly when numerous skeletal metastases serve as a source of cross-fire irradiation to the adjacent marrow. Unlike the classical scenario in which marrow carcinosis predisposes to hematotoxicity, this patient’s toxicity was purely treatment-related, underscoring that the number of cycles must be carefully balanced against the risk of cumulative myelosuppression, even in patients with initially preserved marrow function.

A notable finding in our cohort was transient pain flares in two patients (14.2%) shortly after administration. One patient required discontinuation of treatment due to this event. This phenomenon is well documented with hormonal therapies, such as tamoxifen, and with bone-seeking radionuclides like Radium-223, but is less frequently described in FAPI literature [20].

Pathophysiologically, this flare may reflect an acute inflammatory response within the TME. Rapid binding of the radioligand to high-density FAP receptors on CAFs may trigger localized cytokine release or acute cellular swelling and edema within the confined space of bone metastases or solid masses, temporarily compressing adjacent nerves [21]. Although clinically distressing, a flare reaction often indicates high target engagement. Future protocols may benefit from prophylactic corticosteroids in patients with extensive painful metastases to mitigate this stromal edema [5].

The distinct toxicity and efficacy profiles observed in our cohort broadly align with earlier human experience with FAPI-RLT but reveal pronounced disease-specific differences. Banihashemian et al. reported on four cycles of [¹⁷⁷Lu]Lu-FAPI-2286 (6660–7400 MBq per cycle) in five evaluable sarcoma patients and observed no grade 3/4 adverse events, a partial response by RECIST 1.1 in four of five patients, a 53% reduction in mean primary-tumor volume, and a mean overall survival of 7.8 months from treatment start [19]. In contrast, our breast and GI cancer cohort, treated with a median of only 1.5 cycles, experienced grade 3 anemia in two patients and grade 4 thrombocytopenia in one, with no objective responses and a median overall survival of only 3.25 months. This discrepancy is partly explained by the fundamentally different biology of FAP expression: in sarcomas, FAP is frequently co-expressed on tumor cells, enabling direct malignant-cell kill, whereas in breast and GI carcinomas, FAP is confined to CAFs, so [¹⁷⁷Lu]Lu-FAPI-2286 delivers primarily stromal irradiation that produces palliation but rarely tumor regression. Additionally, the sarcoma patients were less heavily pre-treated and had a better baseline performance status, whereas our patients had extensive skeletal and visceral disease with pre-existing marrow compromise, limiting both the number of cycles that could be administered and the therapeutic window.

Ballal et al. [22] reported meaningful symptomatic relief in ten heavily pre-treated cancer patients receiving [¹⁷⁷Lu]Lu-DOTA.SA.FAPI and [¹⁷⁷Lu]Lu-DOTAGA.(SA.FAPi)₂, without objective radiographic response but with clear palliative benefit. The [¹⁷⁷Lu]Lu-DOTA.SA.FAPi and [¹⁷⁷Lu]Lu-DOTAGA.(SA.FAPi)₂ were well tolerated. No early adverse events occurred after administration. Only one patient with extensive skeletal metastases and pre-existing grade I anemia experienced grade III anemia and grade I thrombocytopenia. No other grade III or IV toxicities were observed.

The comprehensive review by Privé et al. [5] summarized that FAP-targeted RLT is generally well-tolerated, with a low incidence of severe adverse events, though objective response rates remain modest, perhaps especially in tumors where FAP is exclusively stromal. Our safety data—preserved renal and hepatic function, manageable but occasionally severe hematotoxicity in patients with bone-marrow infiltration, and transient pain flares—are consistent with this literature. The palliative pain relief observed in 28.6% of our patients, predominantly those with extensive bone metastases, parallels the symptomatic improvement reported in the sarcoma series and in several FAPI-04/FAPI-46 therapy cohorts, reinforcing the concept that stromal irradiation can alleviate tumor-induced pain even in the absence of RECIST responses.

Another study by Assadi et al. [23] used [¹⁷⁷Lu]Lu-FAPI-46. In this study of 21 heavily pre-treated patients with diverse solid tumors, no grade 3/4 adverse events were observed, except for a single case of grade 3 anemia during concomitant chemotherapy; one patient reported increased bone pain. The median progression-free survival was 3.0 months, and the median overall survival from treatment initiation was 4.0 months at the time of analysis. Of 18 patients, 12 (66.7%) achieved stable disease as best response, and none demonstrated objective remission. Our cohort, treated with the cyclic [¹⁷⁷Lu]Lu-FAPI-2286, showed a broadly comparable safety profile—preserved renal and hepatic function, mild hematological changes—but we encountered more pronounced hematotoxicity (grade 4 thrombocytopenia) and a higher rate of unverifiable radiographic responses, likely reflecting differences in tumor burden, the longer tumor retention time of the cyclic peptide, and the absence of concomitant chemotherapy in most of our patients. The palliative benefit in terms of pain relief (28.6% in our series vs. 33.3% moderate symptom improvement in Assadi et al.) was similar, underscoring that FAPI-RLT, irrespective of the inhibitor scaffold, can provide meaningful symptomatic control even when objective tumor shrinkage is absent.

The findings of the present study should be contextualized within the growing body of evidence on FAPI-targeted RLT in GI malignancies. Several observations from Sweedat et al. [24] are directly pertinent to our results. First, the authors emphasize that GI malignancies — characterized by a highly desmoplastic tumor microenvironment and abundant CAF infiltration — represent a biologically rational and clinically promising target class for FAPI-directed therapy, owing to the dense stromal FAP overexpression that defines this tumor category. This aligns with the rationale underlying patient selection in our cohort, where high FAP expression, confirmed by pre-therapeutic [⁶⁸Ga]Ga-FAPI PET/CT, was an eligibility criterion. Second, Sweedat et al. highlight that ligand-dependent pharmacokinetic differences — including variable tumor retention kinetics across distinct FAPI scaffolds — have important dosimetric implications and may substantially influence both absorbed tumor dose and the therapeutic window. The pharmacokinetic profile of FAP-2286, which features favorable tumor retention relative to earlier-generation FAPI compounds, is relevant in this context and may, in part, explain the disease stabilization observed in a proportion of our patients. Third, the review underscores the heterogeneous response assessment methodologies employed across the published FAPI-RLT literature and advocates for the standardized application of RECIST-based criteria in future prospective studies — a limitation we have also acknowledged in the present work and sought to address more explicitly in the revised manuscript. Finally, Sweedat et al. conclude that available data demonstrate feasibility and an acceptable short-term safety profile for [¹⁷⁷Lu]Lu-FAPI-RLT in GI malignancies. However, evaluating treatment response is challenging for patients undergoing Lu-FAPI-2286, especially those with GI cancer, even though FAPI PET/CT can be a promising tool [7]. No established objective tool exists to assess treatment response, and an objective complete response is rare [24].

Furthermore, we highlight that [¹⁷⁷Lu]Lu-FAPI-2286 has a consistent, class-wide safety profile, but clinical benefit is heavily modulated by tumor histology, FAP compartment (stromal versus tumor-cell), disease burden, and the timing of RLT in the disease trajectory. While the cyclic FAPI-2286 construct offers longer tumor retention than earlier monomers, translating this pharmacokinetic advantage into improved overall survival will likely require selecting patients earlier in the disease course, combination treatments that simultaneously target the unprotected epithelial clone, and personalized dosimetry to avoid dose-limiting hematotoxicity in patients with marrow carcinosis.

When considering the aggregate evidence from Banihashemian et al. [19] (FAPI-2286 in sarcoma), Kuyumcu et al. [25] (FAPI-04), Ballal et al. [22] (DOTA.SA.FAPI), Privé et al. [5] (comprehensive review), and Assadi et al. [23] (FAPI-46), several consistent themes emerge. First, FAP-targeted radioligand therapy is feasible and generally well tolerated, with hematotoxicity the principal dose-limiting concern, especially in patients with pre-existing bone-marrow involvement. Second, objective radiographic responses are uncommon in carcinomas with stromal-restricted FAP expression, whereas sarcomas, which often co-express FAP on tumor cells, appear to derive greater volumetric benefit. Third, palliation of pain is a recurrent observation regardless of the FAPI variant used, indicating that stromal disruption alone can interfere with nociceptive signaling. Fourth, the monomeric inhibitors (FAPI-04, FAPI-46) and the newer cyclic FAPI-2286 share similar organ dosimetry and safety profiles, but the prolonged tumor retention of FAPI-2286, while theoretically advantageous for absorbed dose delivery, may also increase the risk of hematotoxicity in patients with extensive skeletal disease if not guided by personalized dosimetry.

We observed discordance between tumor markers and imaging in specific cases. A patient with rectal cancer showed rising CEA levels despite radiographic qualitative assessment of SD. This uncoupling highlights a fundamental limitation of stromal targeting: FAPI targets CAFs, whereas biomarkers like CEA are produced by epithelial tumor cells. It is biologically plausible that successful stromal suppression could yield stable imaging findings on FAPI-SPECT while the epithelial clone continues to proliferate or dedifferentiate. This suggests that [¹⁷⁷Lu]Lu-FAPI-2286 monotherapy may control the microenvironment but may require combination with cytotoxic agents or molecular therapies to simultaneously ablate the tumor parenchyma.

Strengths, limitations, and future directions

The primary strength of this study is its evaluation of [¹⁷⁷Lu]Lu-FAPI-2286, a next-generation cyclic peptide, rather than the more commonly studied monomeric FAPI-04 or FAPI-46. Our findings provide critical real-world data on the clinical behavior of this radioligand, which theoretically offers higher binding affinity and prolonged tumor retention. Furthermore, this study was conducted in a salvage setting involving heavily pre-treated patients who had exhausted all standard-of-care options, including immunotherapy and chemotherapy. By demonstrating symptomatic relief, specifically pain reduction in bone metastases, in a cohort with no other therapeutic options, this study validates the role of FAP-targeted RLT as a potent palliative tool. Additionally, the identification of the pain flare phenomenon and the discordance between tumor markers and stromal imaging provide novel clinical insights that refine safety monitoring protocols for future FAPI administrations.

We acknowledge several limitations inherent to this pilot compassionate-use study. The cohort size (n = 14) and the inclusion of diverse tumor histologies (breast, colorectal, gastric) limit the statistical power to determine progression-free survival (PFS) or overall survival (OS) for specific subtypes. Because ¹⁷⁷Lu-FAPI RLT is not officially approved by guidelines or the FDA, its use is limited and can be applied only to heavily pretreated patients, who generally present with heterogeneous cancer types. A significant proportion of the cohort (50%) had unverifiable radiographic responses due to rapid clinical deterioration or mortality before follow-up imaging could be performed. This high attrition rate reflects the end-stage nature of the population but introduces a survival bias, potentially skewing efficacy data toward patients with inherently slower-growing disease. We did not perform serial post-treatment dosimetry. Consequently, we could not correlate the absorbed radiation dose in the tumor or critical organs with clinical outcomes or toxicities, such as the Grade 4 thrombocytopenia observed in one patient. As a single-arm observational study, efficacy comparisons against best supportive care are descriptive rather than statistical.

Based on our findings, future research should focus on three strategies: the rapid progression in our cohort suggests that [¹⁷⁷Lu]Lu-FAPI-2286 should be evaluated in earlier lines of therapy, when tumor volume is lower and bone marrow reserve is intact. Administering RLT before the onset of bulky disease may enhance the cross-fire effect and reduce early mortality. The observed discordance between rising tumor markers (epithelial compartment) and stable FAPI imaging (stromal compartment) supports the biological rationale for combination therapies. Future trials should explore combining FAPI-RLT (to deplete the stroma) with cytotoxic chemotherapy or immunotherapy (to target the epithelial clone), potentially overcoming physical barriers to drug delivery. Given the observation of “pain flares” and hematotoxicity, future protocols should incorporate prophylactic corticosteroids for patients with high skeletal tumor burden to mitigate stromal inflammation. Additionally, personalized dosimetry should be mandatory for patients with suspected bone marrow carcinosis to prevent high-grade hematological toxicity.

Conclusion

[¹⁷⁷Lu]Lu-FAPI-2286 demonstrates feasibility and an acceptable safety profile in heavily pre-treated patients with advanced solid tumors. Although objective radiographic responses were modest, the therapy provided meaningful palliative benefit, particularly pain reduction, for a subset of patients. The observed hematotoxicity in patients with marrow involvement and the occurrence of pain flares warrant careful patient monitoring. Future prospective trials should focus on administering FAPI-RLT at earlier disease stages and on exploring combination strategies to target both the tumor stroma and the malignant cells synergistically.

Abbreviations

GI

Gastrointestinal

FAP

Fibroblast activation protein

FAPI

Fibroblast activation protein inhibitor

CTCAE

Common Terminology Criteria for Adverse Events

TME

tumor microenvironment

CAF

cancer-associated fibroblast

RLT

radioligand therapy

ECOG

Eastern Cooperative Oncology Group

AE

adverse events

AST

aspartate transaminase

ALT

alanine transaminase

ALP

alkaline phosphatase

LDH

lactate dehydrogenase

WBC

white blood cell

Hb

hemoglobin

Plt

platelet

CR

complete response

PR

partial response

SD

stable disease

PD

progressive disease

TNBC

triple-negative breast cancer

OS

overall survival

PFS

progression-free survival

Authors’ contributions

Faezeh Rabbani Banoo: Conceptualisation, Data collection, drafting of the manuscript. – Mohammad Hadi Samadi: conceptualisation of the study, data collection, data interpretation, and critical revision. - Sajjad Sadeghpour: study design, methodological oversight, data collection and analysis, and final revision of the manuscript. – Somaye Barashki: patient referral, study design, and clinical and critical revision of the manuscript. - Kamran Aryana: patient referral, clinical expertise, final revision. Ali Emadi Torghabeh: Patient referral, clinical expertise, final revision. - Slaman Soltani: patient referralclinical expertise, final revision. - Ehsan Soltani: patient referral, clinical expertise, final revision - Atena Aghaee: patient referral, clinical expertise, interpretation of nuclear medicine findings, and manuscript revision. All authors read and approved the final version of the manuscript.

Funding

The present study received no funding.

Data availability

All data analysed during this study are included in the manuscript. The data extracted from the Ward dataset are not publicly available due to ethical reasons and are available from the corresponding author upon reasonable request.

Declarations

Ethics approval and consent to participate

This study was approved by the local committee (Mashhad University of Medical Sciences Ethical Committee) in accordance with the ethical code IR.MUMS.MEDICAL.REC.1404.119. The protocol was conducted in accordance with the ethical principles of the Declaration of Helsinki. All patients provided written informed consent to receive the unapproved radiopharmaceutical, acknowledging the experimental nature of the therapy, potential risks, and the collection of their clinical data for research purposes.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.