First-In-Human Results on the Biodistribution, Pharmacokinetics, and Dosimetry of [177Lu]Lu-DOTA.SA.FAPi and [177Lu]Lu-DOTAGA.(SA.FAPi)2

Recently, great interest has been gained regarding fibroblast activation protein (FAP) as an excellent target for theranostics. Several FAP inhibitor molecules such as [68Ga]Ga-labelled FAPI-02, 04, 46, and DOTA.SA.FAPi have been introduced and are highly promising molecular targets from the imaging point of view. FAP inhibitors introduced via bifunctional DOTA and DOTAGA chelators offer the possibility to complex Lutetium-177 due to an additional coordination site, and are suitable for theranostic applications owing to the increased tumor accumulation and prolonged tumor retention time. However, for therapeutic applications, very little has been accomplished, mainly due to residence times of the compounds. In an attempt to develop a promising therapeutic radiopharmaceutical, the present study aimed to evaluate and compare the biodistribution, pharmacokinetics, and dosimetry of [177Lu]Lu-DOTA.SA.FAPi, and [177Lu]Lu-DOTAGA.(SA.FAPi)2 in patients with various cancers. The FAPi agents, [177Lu]Lu-DOTA.SA.FAPi and [177Lu]Lu-DOTAGA.(SA.FAPi)2, were administered in two different groups of patients. Three patients (mean age—50 years) were treated with a median cumulative activity of 2.96 GBq (IQR: 2.2–3 GBq) [177Lu]Lu-DOTA.SA.FAPi and seven (mean age—51 years) were treated with 1.48 GBq (IQR: 0.6–1.5) of [177Lu]Lu-DOTAGA.(SA.FAPi)2. Patients in both the groups underwent serial imaging whole-body planar and SPECT/CT scans that were acquired between 1 h and 168 h post-injection (p.i.). The residence time and absorbed dose estimate in the source organs and tumor were calculated using OLINDA/EXM 2.2 software. Time versus activity graphs were plotted to determine the effective half-life (Te) in the whole body and lesions for both the radiotracers. Physiological uptake of [177Lu]Lu-DOTA.SA.FAPi was observed in the kidneys, colon, pancreas, liver, gall bladder, oral mucosa, lacrimal glands, and urinary bladder contents. Physiological biodistribution of [177Lu]Lu-DOTAGA.(SA.FAPi)2 involved liver, gall bladder, colon, pancreas, kidneys, and urinary bladder contents, lacrimal glands, oral mucosa, and salivary glands. In the [177Lu]Lu-DOTA.SA.FAPi group, the highest absorbed doses were noted in the kidneys (0.618 ± 0.015 Gy/GBq), followed by the colon (right colon: 0.472 Gy/GBq and left colon: 0.430 Gy/GBq). In the [177Lu]Lu-DOTAGA.(SA.FAPi)2 group, the colon received the highest absorbed dose (right colon: 1.160 Gy/GBq and left colon: 2.870 Gy/GBq), and demonstrated a significantly higher mean absorbed dose than [177Lu]Lu-DOTA.SA.FAPi (p < 0.011). [177Lu]Lu-DOTAGA.(SA.FAPi)2 had significantly longer median whole-body Te compared to that of [177Lu]Lu-DOTA.SA.FAPi [46.2 h (IQR: 38.5–70.1) vs. 23.1 h (IQR: 17.8–31.5); p-0.0167]. The Te of tumor lesions was significantly higher for [177Lu]Lu-DOTAGA.(SA.FAPi)2 compared to [177Lu]Lu-DOTA.SA.FAPi [86.6 h (IQR: 34.3–94.6) vs. 14 h (IQR: 12.8–15.5); p-0.0004]. The median absorbed doses to the lesions were 0.603 (IQR: 0.230–1.810) Gy/GBq and 6.70 (IQR: 3.40–49) Gy/GBq dose per cycle in the [177Lu]Lu-DOTA.SA.FAPi, and [177Lu]Lu-DOTAGA.(SA.FAPi)2 groups, respectively. The first clinical dosimetry study demonstrated significantly higher tumor absorbed doses with [177Lu]Lu-DOTAGA.(SA.FAPi)2 compared to [177Lu]Lu-DOTA.SA.FAPi. [177Lu]Lu-DOTAGA.(SA.FAPi)2 is safe and unveiled new frontiers to treat various end-stage cancer patients with a theranostic approach.


Introduction
The tumor microenvironment (TME) plays a crucial role in tumor remodelling and is an important contributor to tumor growth and promoting drug resistance. Within the TME, cancer-associated fibroblasts (CAFs) have a multifaceted function and are major contributors to TME remodelling. The high abundance of CAFs in a wide range of tumors offers important implications to target various cancers.
The fibroblast activation protein (FAPα), a type II transmembrane serine protease, is highly expressed in CAFs. Histopathologic studies reported the prevalence of FAP-positive cancer-associated fibroblasts in~90% of epithelial tumors [1]. The ubiquitous expression of fibroblast activation protein (FAP) makes it an interesting target for imaging and therapy of a wide spectrum of malignancies [1]. FAP promotes tumor growth, proliferation, and angiogenesis [2]. Hence, targeting this protein with several probes, including antibodies, immunoconjugates, and small molecular FAP inhibitors, may be an interesting approach for tumor detection and suppression.
Although FAP imaging is in the early developmental stage, several FAP inhibitor based small molecules (chelator-linker-FAP inhibitor conjugates) have been developed. Mostly, heterocyclic linker units between chelator and inhibitor were introduced, such as piperazine series [3,4] and squaramide based [5][6][7] FAP-inhibitor precursors. All these FAPi precursors exhibit high selectivity and affinities for FAP in the same order of magnitude as the lead structure UAMC1110, as described by van der Veken's group [8,9], and have been developed for diagnostic and therapeutic use ( Figure 1).

Introduction
The tumor microenvironment (TME) plays a crucial role in tumor remodelling and is an important contributor to tumor growth and promoting drug resistance. Within the TME, cancer-associated fibroblasts (CAFs) have a multifaceted function and are major contributors to TME remodelling. The high abundance of CAFs in a wide range of tumors offers important implications to target various cancers.
The fibroblast activation protein (FAPα), a type II transmembrane serine protease, is highly expressed in CAFs. Histopathologic studies reported the prevalence of FAP-positive cancer-associated fibroblasts in ~90% of epithelial tumors [1]. The ubiquitous expression of fibroblast activation protein (FAP) makes it an interesting target for imaging and therapy of a wide spectrum of malignancies [1]. FAP promotes tumor growth, proliferation, and angiogenesis [2]. Hence, targeting this protein with several probes, including antibodies, immunoconjugates, and small molecular FAP inhibitors, may be an interesting approach for tumor detection and suppression.
Although FAP imaging is in the early developmental stage, several FAP inhibitor based small molecules (chelator-linker-FAP inhibitor conjugates) have been developed. Mostly, heterocyclic linker units between chelator and inhibitor were introduced, such as piperazine series [3,4] and squaramide based [5][6][7] FAP-inhibitor precursors. All these FAPi precursors exhibit high selectivity and affinities for FAP in the same order of magnitude as the lead structure UAMC1110, as described by van der Veken's group [8,9], and have been developed for diagnostic and therapeutic use ( Figure 1).  Haberkorn's group reported a series of piperazine-based FAP-inhibitors labelled with the positron emitter gallium-68, which were successfully used for imaging various cancers [3,4], particularly when utilizing the most prominent molecules among their structures such as FAPI-04, FAPI-21, and FAPI-46 [4]. Roesch's group, in collaboration with our group, introduced a modified ligand, keeping the pharmacophore intact as new FAPi PET tracers. The critical subunits constituted the squaramide (SA) linker unit coupled with the DOTA/DATA 5m bifunctional chelators and a FAP inhibitor targeting moiety. Both agents were coupled with generator produced gallium-68 and revealed promising imaging and theranostic benefits on in vitro, preclinical, and clinical studies [5,6]. It is well established now that these FAP inhibitors show expression in various cancers [5][6][7]10]. The monomeric DOTA.SA.FAPi labelled with gallium-68 showed the most favourable properties from the imaging point of view, which includes high tumor-to-background ratios [TBR] and demonstrates a great applicability for the theranostic treatment approach for various cancers.
Our group [7] applied a theranostic approach of [ 68 Ga]Ga guided [ 177 Lu]Lu-DOTA. SA.FAPi therapy in an advanced stage breast cancer (histology status: ER − , PR − , HER2/neu + ) patient who failed multiple lines of treatment, and demonstrated a promising improvement in the quality of life. This radio-ligand therapy concept unveiled a new milestone in precision oncology. However, the findings were preliminary, and the detailed pharmacokinetics and dosimetry data were underway. Visual analysis on PTx-[ 177 Lu]Lu-DOTA.SA.SA.FAPi whole-body scan demonstrated a high tumor affinity, but early washout of the radiotracer, which was completely eliminated by 48 h p.i., was the major drawback of the molecule. However, despite the short-tumor retention time, the patient experienced an improvement in the clinical status.
To overcome this problem, Moon et al. [11] modified the structure and introduced dimeric systems for prolonged tumor retention. Using the SA.FAPi monomer as the base, they developed two homodimeric structures such as DOTA(SA.FAPi) 2 and DOTAGA. (SA.FAPi) 2 ( Figure 1).
The DOTAGA.(SA.FAPi) 2 is based on the monomeric DOTA.SA.FAPi structure, but unlike the monomer, two identical SA.FAPi units are bound to a trifunctional DOTAGA chelator forming a homodimeric system. Additionally, for the possibility of complexing radiometals such as lutetium-177 or actinium-225, at least seven coordinations are required; hence, DOTAGA as a chelator was used in the case of the dimer (Figure 1).
Various derivatives were investigated in in vitro binding assays to FAP, DPPs (prolinespecific enzymes dipeptidyl peptidases), and PREP (prolyl oligopeptidase), and revealed high affinity and protease selectivity to FAP and towards DPPs and PREP (Table 1).  experienced grade III and grade I anaemia and thrombocytopenia, respectively. No other grade III/IV toxicities were noted. Table 3 summarises the median pre-treatment and six months post-treatment hematological, kidney, and liver function parameters that showed that both radiotracers were well endured. On qualitative analysis, maximum normal physiological uptake was observed in the kidneys, followed by the colon/large intestines (ascending, transverse, and descending colon). Other organs included the liver, pancreas, gall bladder, oral mucosa, lacrimal gland, and urinary bladder contents. Time-activity curves were derived by either mono or biexponential curve fitting. [ 177 Lu]Lu-DOTA.SA.FAPi is excreted via both renal excretion and hepatobiliary clearance. A combined wash-in and washout trend of the radiotracer in the kidneys was observed in all patients. The wash-in of the radiotracer in the kidneys was initiated as early as 6 h p.i. and continued up to 144 h after treatment. Pure washout of the trend was observed for lacrimal glands, oral mucosa, liver, and pancreas ( Figure 2).

Safety
Transit of radiotracer in the gut (ascending and descending colon) was first observed at 24 h post-injection and reached its peak uptake at 48 h. At 144 h p.i., a complete washout of the radiotracer was observed from the gut. Figure  Transit of radiotracer in the gut (ascending and descending colon) was first observed at 24 h post-injection and reached its peak uptake at 48 h. At 144 h p.i., a complete washout of the radiotracer was observed from the gut. Figure  Physiological biodistribution of [ 177 Lu]Lu-DOTAGA.(SA.FAPi)2 involved liver, gall bladder, large intestines (transverse, ascending, and descending colon), pancreas, kidneys, urinary bladder contents, and, to a lesser extent, the lacrimal glands, oral mucosa, salivary glands ( Figure 3). Visual analysis revealed the colon as the organ with the highest FAP uptake. The route of excretion was predominantly via biliary followed by renal excretion. A pure washout trend was observed by the kidneys, and a combined wash-in and washout trend of radiotracer was observed by the biliary route and fitted with bi-exponential curves. Kidney excretion was seen as early as 1 h p.i. and continued up to 168 h ( Figure 3).  2 involved liver, gall bladder, large intestines (transverse, ascending, and descending colon), pancreas, kidneys, urinary bladder contents, and, to a lesser extent, the lacrimal glands, oral mucosa, salivary glands ( Figure 3). Visual analysis revealed the colon as the organ with the highest FAP uptake. The route of excretion was predominantly via biliary followed by renal excretion. A pure washout trend was observed by the kidneys, and a combined wash-in and washout trend of radiotracer was observed by the biliary route and fitted with bi-exponential curves. Kidney excretion was seen as early as 1 h p.i. and continued up to 168 h ( Figure 3).
Excretion of radiotracer from the hepatobiliary system/liver commenced at 4 h p.i. and rapidly reduced by 50% at 24 h p.i. The radiotracer concentration in the colon was observed at 24 h p.i. (mean %IA: 14%), showed the maximum uptake at 48 h p.i. (%IA: 17.8%), and washed out to as low as 3% at 168 h post-infusion. Clearance of gut activity by approximately 7.4-fold was observed between 96 h and 168 h. However, the radiotracer concentration in the gut widely varied among the patients depending on the tumor burden, intestinal motility, and excretion. Patients with a high tumor burden received lower absorbed doses to the colon and kidneys due to the "tumor-sink" effect. Excretion of radiotracer from the hepatobiliary system/liver commenced at 4 h p.i. and rapidly reduced by 50% at 24 h p.i. The radiotracer concentration in the colon was observed at 24 h p.i. (mean %IA: 14%), showed the maximum uptake at 48 h p.i. (%IA: 17.8%), and washed out to as low as 3% at 168 h post-infusion. Clearance of gut activity by approximately 7.4-fold was observed between 96 h and 168 h. However, the radiotracer concentration in the gut widely varied among the patients depending on the tumor burden, intestinal motility, and excretion. Patients with a high tumor burden received lower absorbed doses to the colon and kidneys due to the "tumor-sink" effect.

Dosimetry Estimate and Te of Normal Organs
The        (Table 7).

Discussion
The expression of cancer-associated FAP in a broad spectrum of cancers offers an optimal target for various molecular-based FAP inhibitor imaging and therapies [6,7,12].
Based on the synthesis of a potent FAP inhibitor UAMC1110 [8,9], Moon et al. [ Y]Y-FAPI-04 was not conclusive, the patient experienced a reduction in pain with no significant toxicities [3].
From the reports of previous studies and our study on [ 177 Lu]Lu-DOTA.SA.FAPi [3,7], it is evident that the main challenge for the potential therapeutic application of the FAP tracers was to optimize its tumor retention time. To design an ideal radiotracer for theranostic use and deliver a maximum radiation dose to the desired target lesions, the biological half-life of the FAPI agent should match the physical half-life of the radiometal. Small-molecule inhibitors with a shorter-biological half-life could be labelled with shorter physical half-life therapeutic radionuclides such as 90 Y/ 188 Re/ 213 Bi, and similar molecules with longer half-life could be tagged to long-lived therapeutic radionuclides such as 177 Lu/ 225 Ac, etc.
This problem received substantial interest, and an approach to improve tumor affinity as well as tumor retention led to the evolution from monomers to dimeric systems, such as DOTA based homodimeric structures DOTA.(SA.FAPi) 2 and DOTAGA.(SA.FAPi) 2 , by Moon et al. [11]. Unlike monomeric precursors, the bifunctional chelator at the centre is linked to two squaramide linker/target vector units. The coupling of squaramide linker-target vector (FAP inhibitor) to dimers significantly increases tumor uptake, tumor retention, and low background. DOTA.(SA.FAPi) 2 and DOTAGA.(SA.FAPi) 2 were synthesized, tested for stability in vitro, and complexed with gallium-68 and lutetium-177. As evidenced from the reports of Moon et al. [11], the homodimers display very high in vitro affinity for FAP, similar to the monomeric structures (similar low nM IC 50 values). The use of the DOTAGA chelator for the dimeric version allows for the same coordination sites as the DOTA monomer structure, attributing it to the better chelation with heavy radiometals such as 177 Lu, 90 Y, 188 Re, 225 Ac, 213 Bi, etc. It increases the tumor retention time by several folds. Interestingly, the homodimeric structure had significantly increased tumor uptake and retention, with a low background at 24 h p.i. compared to the monomer. To introduce the homodimers from bench to bedside, the systematic clinical trials focusing on head-to-head comparisons of the homodimers addressing the pharmacokinetics in normal organ and tumor lesions are warranted.
Both radiotracers were well tolerated in all patients with minimal toxicities. While the dose-limiting organ with [ 177 Lu]Lu-DOTA.SA.FAPi was the kidney, followed by a colon, the highest estimated absorbed radiation dose by [ 177 Lu]Lu-DOTAGA.(SA.FAPi) 2 dimer was observed in the colon, followed by gall bladder, pancreas, and kidneys.
To achieve a safe limit of 28 Gy [13]  group, one patient with paraganglioma suffered from constipation, with the persistence of activity in the gut even up to 168 h p.i., and hence received a relatively higher dose absorbed to the colon (Figure 4). The physiological uptake in the gut/intestines varied widely across the patients and was majorly dependent on intestinal motility. Efforts to reduce the risk to the colon and reduce the absorption might be facilitated by suggesting that high fatty food to accelerate the washout from the gall bladder and administration of laxatives to accelerate the washout of the radiotracer from the gut may be beneficial. However, the effect of the above methods cannot be deduced from the current study results, and it mandates a proper execution and investigation.
The  the patients and was majorly dependent on intestinal motility. Efforts to reduce to the colon and reduce the absorption might be facilitated by suggesting that hi food to accelerate the washout from the gall bladder and administration of laxa accelerate the washout of the radiotracer from the gut may be beneficial. Howe effect of the above methods cannot be deduced from the current study results, and dates a proper execution and investigation.

Limitations
The stutandardizses certain limitations. The number of patients in the study is small and varied between the groups (three vs. seven). As FAPi based radionuclide therapy is a new treatment, the institute ethics committee approved the [ 177 Lu]Lu-labelled FAPi derivatives as a salvage treatment option. Hence, only the worst prognosis patients who have exhausted all standard line treatment options were recruited and were treated on compassionate grounds, causing a bias in the patient recruitment. The current administered activities to patients were arbitrary, as no data on the dosimetry estimates were available in the literature for any therapeutic FAPi tracer.
Pertaining to the heterogeneity in the type of cancers and difference in the tumor burdens of the patients between the two radiotracer groups, it is not ideal for conducting a head-to-head comparison. The serial time-point of the acquisition was not uniform due to the differences in the pharmacokinetics between the radiotracers. An inherent drawback of planar dosimetry is the overestimation of dose due to the overlap of abdominal organs, but efforts were made to reduce the error by applying appropriate subtraction techniques. The effect of laxatives to promote the early washout of radiotracers and the reduction in radiation burden to the large intestine was out of the scope of this paper.

Future Prospects
Based on the current results, we have initiated a dose-escalation study to evaluate the maximum tolerated absorbed dose to critical organs for [ 177 Lu]Lu-DOTAGA.(SA.FAPi) 2 ; thereby, we expect to achieve the best objective response and minimal toxicities at an optimal dosage of lutetium-177.
From the molecular perspective, we intend to reconstruct/improvise the molecule further, improve its pharmacokinetics to promote minimum uptake in the non-target organs by reducing the percentage of injected activity to the dose-limiting organs (hepatobiliary and large bowels), and enhance tumor internalization and achieve greater "tumor-sink" effect.

Patient Recruitment
The study was duly approved by the ethics committee of the All India Institute of Medical Sciences, New Delhi as a salvage treatment option and for patients who have Patients who received prior anti-cancer therapy within the previous four weeks, patients with Hb < 9 g/dL, leukocyte counts less than 4.0 × 10 9 /L, platelet counts less than 75,000 per mL, inadequate liver function parameters, and serum creatinine > 1.2 mg/dL were excluded from the study.
The study was first initiated using [ 177 Lu]Lu-DOTA.SA.FAPi, but after the preliminary qualitative results of serial imaging, we observed low radiotracer retention at about one to two days p. Patients were positioned in a supine position, and an initial scout was acquired, followed by a diagnostic dose CT with 300-350 mAs, 120 kVp, slice thickness 5 mm, and pitch 1 and PET acquisition with 2 min per bed.
The images were subjected to dead-time, random, and scatter correction. The PET image reconstruction was performed using an ordered subset expectation maximization algorithm (OSEM) (21 subsets, 3 iterations). All images were processed and analyzed on the GE Xeleris workstation.  3 , which was obtained from BRIT, India, in sodium acetate buffer, pH 4, in 0.01 M supra pure HCl. The radiolabelled solution was heated at 95 • C for 30 min, followed by purification through Sep-Pak C18 light cartridge and eluted with 50% ethanol. Radiochemical quality control was carried out using the instant thin-layer chromatography method with sodium citrate buffer as the solvent, and radiolabelled products with 95% to 98% purity were administered. The planar acquisition of whole-body scans was performed using a dual-headed gamma camera (GE, Discovery NM/CT 670). The camera was equipped with a highenergy general-purpose (HEGP) parallel-hole collimator, and the energy peak was centred at 113 keV and 208 keV with a 10% window width. Dual-energy scatter corrections were applied at 90 KeV and 170 keV with a window width of 10%. Serial whole-body emission scans were performed at 1 (pre-void), 6, 24, 48, and 144 hours (h) after treatment for the Similarly, SPECT/CT scans of the abdomen and the lesions were acquired in both the radiotracer groups at serial time points, but were mainly used to demarcate the overlapping gut and kidney activity and to calculate the volume of the tumor. SPECT/CT acquisition parameters included a total angular range of 360 degrees, an angle view of 6 degrees, acquired at 25 s per view, and a matrix size of 512 × 512.

Image Analysis
In the dosimetry analysis, salivary glands, kidneys, pancreas, liver, gall bladder, right colon, left colon, tumor lesions, and whole body were included for dose calculation. The first whole-body image post-injection before voiding was considered to include 100% of injected activity. Background counts were obtained from the thigh region. For overlapping organs such as the right kidney which had overlapping intestinal uptake, the counts were considered to the left kidney. The corresponding time-point post-therapy single photon emission computed tomography (PTx-SPECT/CT) scans were also referred to prevent overlap. Background correction of lesion counts was performed by subtracting counts in background ROI of the similar area drawn close to the lesions.
Finally, attenuated, background, and scatter corrected percentage injected activity (%IA) in each source organ including salivary glands (parotid and submandibular glands), kidney, liver, gall bladder, pancreas, right and left colon, and the tumor was calculated was calculated according to the Equations (1) and (2).
where %IA uncorr : Uncorrected percentage of injected activity; Ct ROI/pixel : counts/pixel in a region of interest; Ct WB/pixel : counts in the whole-body image.
%I A Corr = Ct ROI /pixel Ct WB /pixel × DF × 100 (2) where %IA Corr : Corrected percentage of injected activity (corrected with decay factor); Ct ROI/pixel : counts/pixel in the region of interest; Ct WB/pixel : counts/pixel in whole-body image; DF: decay factor.

Internal Dose Estimation
The percentage injected activities against time were entered in the kinetic input model of the OLINDA/EXM v2.2 software to calculate the area under the curve that represented the number of disintegrations per injected activity, residence time, or cumulative activity in each source organ or time-integrated activity coefficient. The residence times were input to the ICRP-89 female and male models to derive absorbed doses of organs and whole-body.

Tumor Dosimetry
For the tumor dosimetry, a sphere model implemented within OLINDA/EXM v2.2 was used. For each considered lesion, the volume was evaluated on pre-therapy [ 68 Ga]Ga-DOTA.SA.FAPi PET/CT and PTx SPECT-CT of the area of interest using the commercially available workstation (GE Xeleris).
For the estimation of tumor absorbed dose, the dose equation based on the MIRD formalism is expressed below [17,18] [Equation (3)].
Here, τ is the residence time, ∼ A is the cumulated activity, A 0 is the patient's administered activity, and S is the mean absorbed dose per unit cumulated activity.
Finally, the residence times of source organs and tumors were entered in the adult female or male ICRP 89 model for normal organs and the sphere model, respectively, that derived the organ and whole-body absorbed doses in terms of or Gy/GBq. The effective half-lives (T e ) of various organs and tumors were generated using GraphPad Prism software (v9.1).

Safety
Safety was assessed by dosimetry and adverse events assessment according to the National Cancer Institute's Common Toxicity Criteria (NCI-CTCAE) version 5.0.

Statistical Analysis
The D'Agostino Pearson test was used to check for the normal distribution of data. Based on the distribution, summary statistics were obtained in terms of mean, median, standard deviation (SD), range, and interquartile range (IQR) were calculated for all continuous variables based on the distribution of data. The Mann-Whitney test for independent samples was used to compare the organ, tumor absorbed doses, and the Te between the radiotracers. p-value < 0.05 was considered statistically significant. Statistical analysis was performed with MedCalc statistical software version 12.