📄 medicine

Diagnostic Accuracy of Prostate-Specific Membrane Antigen PET/CT for Pelvic Lymph Node Staging in Intermediate-to-High-Risk Prostate Cancer: A Systematic Review and Meta-Analysis

This systematic review and meta-analysis of 30 studies involving 5,927 patients demonstrates that while PSMA PET/CT offers high and reproducible specificity for pelvic lymph node staging in intermediate-to-high-risk prostate cancer, its moderate and heterogeneous sensitivity limits its ability to rule out micrometastatic disease, suggesting it should be used to confirm nodal involvement for treatment intensification rather than as a standalone tool to exclude surgery.

Original authors: Bahareh Mehdikhani, Reza Shahrokhi Damavand, Ehsan Zolfi, Nima Narimani, Mohammad Mehdi Atarod, Hadi Maleki, Saman Farshid, Alireza Negahi, Kazem Aghili, Hossein Neamatzadeh

Published 2026-07-31

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Bahareh Mehdikhani, Reza Shahrokhi Damavand, Ehsan Zolfi, Nima Narimani, Mohammad Mehdi Atarod, Hadi Maleki, Saman Farshid, Alireza Negahi, Kazem Aghili, Hossein Neamatzadeh

Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine you are a detective trying to solve a mystery inside a bustling city. The city is the human body, and the mystery is a hidden gang of troublemakers called cancer cells. In a specific type of cancer called prostate cancer, these troublemakers sometimes try to sneak out of their original neighborhood and hide in the nearby lymph nodes, which act like security checkpoints. If the police (doctors) can find these hidden gangs early, they can choose the right strategy to stop them. But here's the tricky part: the troublemakers are often tiny, like specks of dust, and the old security cameras (standard CT or MRI scans) are too blurry to see them unless they have grown into big, obvious lumps.

Enter a new, super-powered spy tool called PSMA PET/CT. Think of this as a high-tech drone equipped with a special "glow-in-the-dark" tracker. Prostate cancer cells are famous for wearing a specific uniform called PSMA. This drone is designed to light up whenever it sees that uniform, making the hidden troublemakers glow so the police can spot them. Doctors are excited because this tool seems to be changing the game, but they need to know one crucial thing: Is it perfect? Can it find every single hidden gang, or does it sometimes miss the tiny ones? If it misses them, the police might think the coast is clear and let the troublemakers escape, which could be dangerous.

This paper is a massive investigation into how good this super-powered drone really is. The researchers didn't just look at one or two cases; they gathered data from 30 different studies involving nearly 6,000 patients. They wanted to see how often the drone actually found the hidden cancer in the lymph nodes compared to the "gold standard" proof: taking the nodes out and looking at them under a microscope in a lab.

Here is what they found, and it's a bit of a "good news, bad news" story.

The Good News: The Drone is a Master at Saying "No"
When the drone says, "I see cancer here!" it is almost always right. The study found that the tool has a 93.2% success rate at correctly identifying when cancer is present. This is like a metal detector that beeps only when it finds a real coin, not a piece of trash. If the scan lights up, doctors can be very confident that the cancer has spread, and they can immediately switch to a stronger treatment plan to fight it. The tool is incredibly reliable at confirming the bad news.

The Bad News: The Drone Sometimes Misses the Tiny Sneakers
However, when the drone says, "I don't see anything," it isn't always telling the whole truth. The study found that the tool only catches about 56.9% of the hidden cancer cases. This means that if a patient has a "clean" scan, there is still a real chance (about 43%) that tiny, invisible gangs of cancer cells are hiding in the lymph nodes, waiting to cause trouble later. The drone is great at spotting the big, obvious troublemakers, but it often misses the tiny ones that are smaller than a grain of sand (less than 5 millimeters).

Why the Numbers Vary
The researchers also discovered that the drone's performance isn't the same in every situation. It depends on a few things:

The Type of Drone: Some versions of the scanner (using a specific chemical called 68Ga-PSMA-11) were better at finding the troublemakers (65.1% success) than others (like 18F-DCFPyL, which only found 34.0%).
The Investigation Method: In some studies, the researchers only checked the lymph nodes with a microscope if the drone had already spotted something. This made the drone look much better than it really was (showing 81.7% success) because they didn't check the "clean" areas where the tiny troublemakers might have been hiding. When they checked everyone regardless of what the drone saw, the success rate dropped to a more honest 49.4%.
The Patient's Clues: The higher the level of a protein called PSA in the patient's blood, the better the drone seemed to perform.

The Bottom Line
The paper concludes that this super-powered drone is an excellent tool for confirming a problem, but a risky tool for ruling one out. You can trust it when it says "Yes, the cancer is there," but you cannot fully trust it when it says "No, everything is fine."

Because of this, the researchers suggest that doctors shouldn't rely on the scan alone to decide whether to skip a surgery called an "extended pelvic lymph node dissection" (which is basically a thorough search of the lymph nodes). Even if the drone sees nothing, if the patient has a high risk of cancer based on other factors, doctors should still perform the thorough search to make sure they aren't missing those tiny, invisible gangs. The scan is a powerful helper, but it hasn't replaced the need for a careful, hands-on investigation just yet.

Technical Summary: Diagnostic Accuracy of PSMA PET/CT for Pelvic Lymph Node Staging in Intermediate-to-High-Risk Prostate Cancer

Problem Statement
Accurate staging of pelvic lymph nodes (N1 disease) is critical for treatment selection in intermediate-to-high-risk prostate cancer, as it dictates the extension of radiation fields, the addition of androgen-deprivation therapy, or the modification of surgical lymphadenectomy. While extended pelvic lymph node dissection (ePLND) remains the histopathologic reference standard, it carries significant morbidity. Conventional imaging (CT/MRI) lacks sensitivity for small-volume nodal deposits (<5 mm), often missing micrometastases. Although Prostate-Specific Membrane Antigen (PSMA) PET/CT has been rapidly adopted, its per-patient diagnostic accuracy against a histopathologic reference standard—and the specific factors modulating this accuracy—remains incompletely defined. Previous syntheses have been limited by heterogeneous populations (including biochemical recurrence), composite reference standards, or mixed analysis levels (per-lesion vs. per-patient), leading to biased estimates for primary staging.

Methodology
This study is a systematic review and meta-analysis conducted in accordance with PRISMA-DTA 2020 guidelines and a prospectively registered protocol (PROSPERO).

Search Strategy: Five bibliographic databases (MEDLINE, Embase, CENTRAL, Web of Science, Scopus) and grey literature sources were searched from inception to June 8, 2026.
Eligibility Criteria: The analysis was restricted to studies evaluating PSMA PET/CT for primary per-patient N1 staging in treatment-naïve intermediate-to-high-risk prostate cancer. Crucially, the reference standard was required to be histopathologic examination of ePLND specimens. Studies involving biochemical recurrence, composite/imaging-only reference standards, or per-lesion analyses without patient-level data were excluded.
Data Extraction: Two reviewers independently extracted data to construct per-patient 2×2 contingency tables. Discrepancies were resolved via adjudication.
Statistical Analysis: A bivariate random-effects model (Reitsma et al.) was used to jointly model logit-sensitivity and logit-specificity, accounting for between-study heterogeneity and negative correlation. Hierarchical SROC curves (Rutter and Gatsonis) and the Moses–Shapiro–Littenberg model were derived.
Heterogeneity and Bias: Heterogeneity was quantified using $I^2$ and $\tau^2$ . Prespecified subgroup analyses, meta-regression (assessing covariates like tracer type, verification status, and PSA), 14 sensitivity analyses, leave-one-out diagnostics, and publication bias assessments (Deeks' funnel plot, Egger's test) were performed.

Key Contributions

Strict Eligibility: Unlike prior reviews, this analysis isolates the specific clinical scenario of primary staging in treatment-naïve patients using histopathologic ePLND as the sole reference standard, minimizing verification bias.
Comprehensive Bias Assessment: The study explicitly quantifies the impact of partial verification (work-up bias), tracer choice, and study design on diagnostic accuracy.
Prediction Intervals: The analysis provides 95% prediction intervals for sensitivity and specificity, offering a realistic estimate of performance in future clinical settings rather than just a pooled mean.
Robustness: Extensive sensitivity analyses (including exclusion of reconstructed data, small studies, and low-quality studies) confirm the stability of the findings.

Results

Study Characteristics: 30 studies comprising 5,927 patients were included. The pooled nodal prevalence was 25.5%.
Pooled Diagnostic Accuracy:
- Sensitivity: 56.9% (95% CI: 48.3–65.0%; $I^2$ = 86.9%).
- Specificity: 93.2% (95% CI: 92.1–94.2%; $I^2$ = 26.0%).
- Diagnostic Odds Ratio (DOR): 19.03 (95% CI: 12.82–28.23).
- Likelihood Ratios: Positive LR = 7.44; Negative LR = 0.50.
- SROC: Area under the curve (AUC) = 0.90; Q* index = 0.895.
Prediction Intervals: The sensitivity prediction interval was wide (20.2–87.2%), indicating high variability in future study performance. In contrast, the specificity prediction interval was narrow (89.8–95.6%), suggesting specificity is a reproducible, intrinsic property of the test.
Subgroup and Meta-Regression Findings:
- Verification Bias: Studies with partial verification (where only PET-positive patients underwent ePLND) reported significantly higher sensitivity (81.7%) compared to completely verified cohorts (49.4%). Meta-regression identified partial verification as a significant independent predictor of sensitivity ( $p=0.007$ ).
- Tracer Performance: $^{68}$ Ga-PSMA-11 demonstrated the highest sensitivity (65.1%), while $^{18}$ F-DCFPyL showed the lowest (34.0%). Specificity remained uniform across tracers.
- Other Predictors: Median PSA ( $p=0.017$ ) and prior conventional imaging ( $p=0.006$ ) were significant independent predictors of sensitivity.
- Study Quality: Restricting analysis to high-quality (QUADAS-2 low risk) studies yielded a sensitivity of 45.7%, suggesting lower-quality studies tend to overestimate sensitivity.
Publication Bias: Significant small-study effects were detected (Deeks' $p=0.0009$ ), with trim-and-fill analysis suggesting an overestimation of sensitivity in the published literature.

Significance and Claims
The paper concludes that PSMA PET/CT functions as a high-specificity, moderate-sensitivity tool for pelvic N1 staging.

Rule-In Capability: A positive scan reliably confirms nodal disease (high specificity), raising the post-test probability of N1 disease from a baseline of ~25% to ~72%. This supports treatment intensification (e.g., extended radiation fields, systemic therapy).
Rule-Out Limitation: A negative scan cannot exclude micrometastatic involvement. With a sensitivity of roughly 50% (and as low as 43–49% in rigorously verified cohorts), approximately one in seven patients with a negative scan still harbors histologic nodal disease.
Clinical Guidance: The authors assert that a negative PSMA PET/CT should not be used in isolation to withhold extended pelvic lymph node dissection (ePLND) in patients with high nomogram-predicted risk. Instead, the scan should be integrated with validated nomograms and selective ePLND.
Physical Limitations: The moderate sensitivity is attributed to the physical detection threshold of current scanners, which routinely misses sub-5-mm micrometastases, a limitation not resolved by current spatial resolution improvements.

The study emphasizes that while PSMA PET/CT is a transformative technology for primary staging, its utility is delimited by its inability to detect small-volume disease, necessitating a cautious approach to surgical decision-making that avoids reliance on the scan alone for ruling out nodal metastasis.

More like this