ABSTRACT Radiopharmaceutical therapy (RPT) that delivers targeted cytotoxic radiation to cancer cells is typically administered on a fixed activity basis, without personalization, in stark contrast to external beam radiotherapy. 177Lu-labelled prostate-specific membrane antigen (PSMA) targeted radioligand therapy (RLT) is a promising RPT for metastatic castrate resistant prostate cancer (mCRPC) that was recently FDA approved as a fixed activity treatment (7.4 GBq/cycle delivered over 6 cycles). This standardized approach fails to take advantage of unique properties inherent to 177Lu: in addition to therapeutic beta particles, 177Lu emits gamma-rays imageable by SPECT, allowing precise absorbed dose (AD) estimation to plan subsequent cycle/s. Additionally, pre-therapy PSMA-PET imaging can potentially be used to predict ADs before initiating RLT. Furthermore, preliminary data demonstrate an important finding: AD deposition per injected activity (Gy/GBq) is substantially higher in the first cycle for tumor, while it remains relatively constant across cycles for organs, motivating consideration of administering higher activity in the first cycle. While standard 177Lu-PSMA RLT has demonstrated a favorable toxicity profile, the AD to critical organs such as bone marrow (BM) needs to be considered when escalating dosage. However, BM dosimetry in mCRPC is complex because tumor infiltration of spongiosa is common. Although the current approach to RLT has led to promising early response rates in mCRPC, the complete response rate and overall survival remain poor; this raises the question of whether dosage escalation with patient-specific dosimetry may be key to more optimal outcomes, and provides the rationale for our study. Our long-term goal is to achieve durable responses with RLT in mCRPC. To advance this goal, the objectives of the current application are to develop the tools and predictive models that will provide the capacity for dosimetry- guided treatment, and to evaluate the feasibility of dosage escalation. For the proposed developments, except the clinical trial, imaging data from 50 patients who are undergoing standard RLT will be used. 1) In Aim 1, a novel method for BM dosimetry will be developed by coupling macro/micro Monte Carlo modeling with images of lesion and marrow distribution in spongiosa, 2) in Aim 2 models will be built to predict AD from pre-therapy quantitative 68Ga-PSMA PET, tumor AD vs. response, and normal organ/BM AD vs. toxicity using biomarkers as covariates and 3) in Aim 3, tracer kinetic modeling will be performed with dynamic PET to determine optimal timing of escalated dosage delivery to avoid PSMA-receptor saturation that may impede benefits of escalation. This will be followed by a Phase I clinical trial for cycle 1 dosage escalation in 20 patients with low risk for hematotoxicity with the primary endpoint of safety; this will identify the recommended dosage for a future Phase II/III trial powered to evaluate efficac...