When we talk to nuclear medicine departments about implementing patient-specific dosimetry, the conversation almost always arrives at the same inflection point: "Our camera can do quantitative SPECT, but we have not set it up yet." That answer is more common than you might expect, even at centers that have been actively delivering PRRT for several years.
The gap between "hardware capable" and "clinically active" for quantitative SPECT is primarily a workflow and configuration problem, not a technology problem. This post describes the practical steps to close that gap in a standard nuclear medicine department.
What Quantitative SPECT Actually Means
Standard clinical SPECT produces images in which pixel values are proportional to regional radiotracer concentration in arbitrary units. The proportionality constant depends on camera sensitivity, reconstruction parameters, collimator geometry, and image corrections applied during reconstruction. You cannot compare absolute pixel values between different patients, different scan sessions, or different cameras without knowing that constant.
Quantitative SPECT, specifically attenuation-corrected SPECT with a camera-specific sensitivity calibration, produces images in which pixel values are in absolute units: typically Bq/mL or kBq/mL. That absolute calibration is what enables you to measure actual radiotracer activity in a defined organ or tumor volume, which is the prerequisite for dosimetry.
Most SPECT/CT systems from the major vendors have supported iterative reconstruction with CT-based attenuation correction for at least ten years. Quantitative reconstruction modes that output calibrated absolute units have been available on standard clinical platforms for five to seven years. The clinical barrier has not been hardware availability; it has been the operational steps needed to establish and maintain the absolute calibration.
Camera Sensitivity Calibration: The Essential First Step
Before you can produce quantitative images, you need a camera-specific calibration factor that translates reconstructed pixel values into absolute activity. This is the system sensitivity, expressed in units of counts per second per megabecquerel, measured under the same acquisition and reconstruction conditions you will use clinically.
The calibration measurement is not technically complex. You prepare a cylindrical phantom with a known activity of the relevant radionuclide (typically Lu-177 for PRRT dosimetry) at a concentration verified against a dose calibrator that has been calibrated to a traceable reference standard. You acquire a SPECT/CT scan of the phantom using your clinical acquisition protocol and apply the same reconstruction pipeline you will use for patients. From the reconstructed image, you extract the total measured activity in the phantom volume and divide by the reference activity to get the calibration factor.
The critical operational discipline is that this calibration must be repeated whenever any element of the acquisition or reconstruction chain changes: collimator replacement, software update to the reconstruction engine, new reconstruction preset, or significant hardware servicing. A calibration that was valid under one reconstruction version may not be valid under another. Many departments do not have a written protocol for recalibration triggers, which means calibration validity can drift silently after routine camera maintenance.
Reconstruction Protocol Standardization
Quantitative accuracy is sensitive to the reconstruction parameters. Ordered subset expectation maximization (OSEM) reconstruction with CT-based attenuation correction, energy window-based scatter correction, and resolution recovery post-filtering represents current clinical standard practice. The specific number of OSEM iterations and subsets, the resolution recovery kernel, and the post-filter settings affect both image quality and quantitative accuracy.
Within a theranostics program, the reconstruction preset used for dosimetry acquisitions should be locked and version-controlled. This means: named in the acquisition protocol document, saved as a named preset on the camera workstation, and not modified without a corresponding recalibration and documentation update. Using different reconstruction settings for the first and third post-treatment SPECT acquisitions in the same patient introduces a systematic error into the time-activity curve fit that cannot be corrected post hoc.
Most departments we work with through YSOTOPE have not formalized this protocol locking. The reconstruction preset exists, but it may have been modified at some point to improve image appearance for diagnostic reads without being flagged as a change that affects dosimetry calibration. Formalizing the separation between diagnostic reconstruction presets and dosimetry reconstruction presets is one of the first operational changes we recommend.
Acquisition Timing and Serial Scan Logistics
PRRT dosimetry requires serial SPECT imaging at multiple time points after administration. The minimum useful dataset for fitting a mono-exponential clearance model is two time points; three time points substantially improve curve fit reliability and allow detection of bi-exponential behavior. The commonly used time points for post-Lu-177 dosimetry are 24 hours and 168 hours (one week) post-injection, sometimes with an additional 96-hour acquisition.
The logistics of serial imaging create real scheduling challenges. Patients receiving PRRT travel to the treating center for administration and then need to return for the 168-hour scan one week later. For patients traveling from other cities or countries, the 168-hour scan can be acquired at a local gamma camera facility if a remote data transfer and standardized reconstruction agreement is in place. That kind of inter-institutional calibration cross-check is technically feasible but requires explicit setup and quality assurance.
One practical trade-off we have seen departments make: acquiring only the 24-hour time point for later cycles in a multi-cycle series when the patient demonstrates stable clearance kinetics from earlier cycles. This is not equivalent to full cycle dosimetry, but it is a reasonable pragmatic adaptation when patient travel burden is limiting adherence to the imaging protocol. The dosimetry report should clearly document which time points were available for each cycle.
DICOM Data Transfer and Reconstruction Verification
For dosimetry software to access the quantitative image data, you need a reliable DICOM transfer from the camera workstation to the dosimetry analysis platform. Standard DICOM SPECT data does not have a mandatory field for the absolute calibration factor; different vendors store this information in private tags or in separate calibration DICOM objects. Any dosimetry pipeline must handle this vendor heterogeneity explicitly.
We verify calibration factor retrieval as part of every scan ingestion in the YSOTOPE platform. If the calibration factor cannot be read from the DICOM metadata, the scan is flagged for manual review rather than proceeding with unverified pixel value assumptions. This verification step catches a class of error (calibration metadata loss during DICOM transfer or PACS archiving) that would otherwise silently corrupt downstream dose calculations.
Building the Operational Case for Dosimetry in Your Department
The clinical benefit of dosimetry-guided PRRT is increasingly supported by the literature. Limiting renal dose to established constraints while monitoring tumor dose accumulation across cycles provides a rational framework for deciding on activity per cycle, total cycle number, and when to consider treatment modification.
The operational argument is also becoming clearer. The equipment required, namely modern SPECT/CT, a dose calibrator with traceable calibration, and a dosimetry software platform, is present in most PRRT-active centers. The remaining investment is in staff training, protocol documentation, and workflow integration. For a department delivering 50 or more PRRT treatments per year, the incremental staff time per patient for dosimetry can typically be covered by a clinical physicist or appropriately trained nuclear medicine technologist with dedicated sessions built into the weekly schedule.
We are not suggesting that full multi-point dosimetry is operationally straightforward to implement from scratch. It is not. But neither is it the research-grade undertaking it was a decade ago. The barrier today is mostly organizational rather than technical, and organizational barriers respond to clear workflows, written protocols, and small teams willing to treat dosimetry as a clinical standard rather than an optional add-on.