The imaging technology at the center of the study is gallium-68 labeled prostate-specific membrane antigen PET/CT, written as [68Ga]Ga-PSMA PET/CT. PSMA is a protein that sits on the surface of prostate cancer cells, and it becomes far more abundant as the disease grows more aggressive. By attaching a radioactive tracer to a molecule that binds PSMA, clinicians can make prostate cancer cells glow on a PET scanner, even when they have seeded only a handful of distant sites. The PET component measures the radiation signal, while the CT component provides anatomical context, so a suspicious hotspot can be matched to a specific bone. The intensity of the signal at the hottest point of a lesion is quantified as the maximum standardized uptake value, or SUVmax, a number that reflects how avidly that tissue grabs the tracer.

That SUVmax number is where the diagnostic dilemma begins. A bone lesion with modest tracer uptake might be a genuine early metastasis, or it might be a benign process such as degenerative change, inflammation, or a healing fracture. Radiologists must draw a line somewhere: above the line, the lesion is called metastatic; below it, benign. Draw the line high, and you risk missing real disease. Draw it low, and benign hotspots get mislabeled as cancer. The Turkish team, led by Mustafa Erol and colleagues, compared two commonly used thresholds head to head in the same patients: a fixed cutoff of SUVmax greater than 3.5, and a relative cutoff defined as SUVmax greater than 1.5 times the activity measured in the blood pool, which adjusts for each patient’s individual background signal.

The study was retrospective and enrolled 108 patients with untreated prostate cancer who had undergone both [68Ga]Ga-PSMA PET/CT and 99mTc-MDP planar bone scintigraphy within 30 days of each other, all for initial staging. To avoid circular reasoning, the researchers did not judge either scan against itself. Instead, they built a best value comparator, or BVC, for each patient and each lesion, combining clinical data, biochemical markers such as prostate-specific antigen, and follow-up imaging to establish the ground truth. Bone metastases were ultimately confirmed in 38 of the 108 patients, or 35.1 percent, a proportion typical of newly diagnosed intermediate- and high-risk disease.

The results were striking at both the patient and the lesion level. Using the stricter SUVmax greater than 3.5 threshold, PSMA PET/CT achieved 97.4 percent sensitivity, 98.6 percent specificity, 97.4 percent positive predictive value, and 98.6 percent negative predictive value. In plain terms, it caught nearly every patient with bone metastases while almost never raising a false alarm. When the researchers loosened the criterion to the relative blood pool threshold, sensitivity and negative predictive value rose to a perfect 100 percent, meaning no metastatic patient went undetected. The price was a modest drop in specificity to 94.3 percent, driven by benign tissue taking up the tracer. Conventional bone scintigraphy, by comparison, managed only 84.2 percent sensitivity and 61.4 percent specificity, meaning it missed roughly one in six patients with bone spread and misclassified a large share of patients without it.

The lesion-level analysis made the contrast even sharper. Among 147 metastases confirmed by the best value comparator, PSMA PET/CT with the blood pool threshold detected every single one, while bone scintigraphy identified only 109 lesions, or 74.1 percent, leaving 38 metastases invisible. The lesions the bone scan missed were disproportionately oligometastatic deposits, meaning small numbers of distant lesions, and osteolytic lesions, which destroy bone rather than provoking the dense new bone formation that technetium tracers highlight. That mechanistic detail explains the old test’s blind spot: 99mTc-MDP accumulates where bone is remodeling in response to tumor, so a lesion that quietly erodes bone without triggering a strong repair response simply does not light up.

What about the threshold question itself? When the two SUVmax criteria were compared directly using exact McNemar testing, the differences were not statistically significant: p equaled 1.00 for sensitivity and 0.25 for specificity. The blood pool threshold offered slightly better sensitivity, the fixed 3.5 cutoff slightly better specificity, but neither advantage reached the level of statistical significance. For clinicians, that is arguably good news, because it suggests PSMA PET/CT’s performance is robust across reasonable interpretation criteria, and radiologists have some latitude in how aggressively they call borderline lesions without dramatically changing the test’s overall accuracy.

The clinical implications reach beyond diagnostic bragging rights. Staging accuracy determines whether a patient receives localized therapy, androgen deprivation combined with radiation, or systemic treatment, and increasingly whether oligometastatic disease is eligible for metastasis-directed therapies. A bone scan that misses 38 of 147 metastatic lesions could leave patients under-staged and under-treated, while its modest specificity can trigger unnecessary anxiety and additional workups. The study’s findings align with and reinforce current clinical guidelines that already favor PSMA PET/CT over planar bone scintigraphy for staging prostate cancer, providing the kind of threshold-specific, head-to-head evidence that guideline committees need when refining recommendations.

The authors are careful about the study’s limits, as any rigorous retrospective single-center analysis must be. The patient cohort, the scanner technology, and the interpretation practices at one institution may not generalize everywhere, and the best value comparator, while carefully constructed, is itself imperfect because no imaging test can achieve perfect ground truth in living patients. The researchers also disclosed that an AI language tool was used only to polish the manuscript’s English, not for study design, data analysis, or interpretation, with the authors retaining full responsibility for the content. Those caveats notwithstanding, the core message stands on solid ground: for detecting bone metastases in untreated prostate cancer, [68Ga]Ga-PSMA PET/CT is not merely an alternative to the traditional bone scan, it is decisively better, and its advantage holds whether radiologists use a fixed SUVmax cutoff or one calibrated to each patient’s blood pool activity. As PSMA-targeted imaging becomes more widely available and less expensive, the planar bone scan’s role in prostate cancer staging looks increasingly like a relic worth retiring.

Subject of Research: Diagnostic accuracy of PSMA PET/CT versus bone scintigraphy for detecting bone metastases in prostate cancer

Article Title: Impact of different SUVmax thresholds on diagnostic accuracy of [68Ga]Ga-PSMA PET/CT versus 99mTc-MDP planar bone scintigraphy in prostate cancer

Article References: Erol, M., Şen, A. E., Tombak, H. A., Şahin, Ö., & Kaya, B. (2026). Impact of different SUVmax thresholds on diagnostic accuracy of [68Ga]Ga-PSMA PET/CT versus 99mTc-MDP planar bone scintigraphy in prostate cancer. BMC Medical Imaging. https://doi.org/10.1186/s12880-026-02902-2

Image Credits: AI Generated

DOI: 10.1186/s12880-026-02902-2

Keywords: prostate cancer, bone metastases, PSMA PET/CT, bone scintigraphy, SUVmax threshold, 99mTc-MDP, gallium-68, diagnostic accuracy, cancer staging, nuclear medicine, oligometastatic disease, medical imaging