The Paris System for reporting Urinary Cytology

Urinary cytology has been widely utilized as a noninvasive and cost-effective screening and diagnostic test for patients with bladder cancer. However, the major limitation of this test is its inability to detect low-grade tumors, often resulting in a high rate of indeterminate (atypical) diagnosis. Dr. Tian explains in this video that by vigorously implementing "The Paris System" for reporting urine cytology, how the lab can achieve better quality and minimize borderline diagnosis.

[@lamwlw]

Transcript:
"
We will discuss the diagnostic criteria in each category in detail. First, we will discuss adequacy in urinary cytology. The following conditions meet the criteria for “Unsatisfactory for Evaluation”:
•Acellular or virtually acellular sample
•Less than 10 urothelial cells for a voided urine or less than 15 cells for an instrumented urine
•Presence of only keratinizing squamous cells
•Specimen contains >75% obscuring debris, inflammation, or lubricant
•Presence of any atypical or malignant cells is adequate for evaluation.
•For voided urine: >30 ml, contains any urothelial cells; for instrumented urine: ≥2 urothelial cells/HPF or ≥200 cells.
NHGUC: negative for high-grade urothelial carcinoma. This includes a variety of situations that previously could be in the “Atypical” category. It includes benign urothelial, glandular, and squamous cells; fragments; clusters; changes with urolithiasis; viral cytopathic effect; post-therapy effects; diversion urine, etc.
It is important to know this category well, as some are close mimickers of atypical urothelial cells. We will discuss this category in detail.
Umbrella cells can be mono-, bi-, or multi-nucleated. Cytoplasm is dense or vacuolated. Cell membrane is sharply demarcated, generally with one flat edge. Multi-nucleation is called “endomitosis” and is caused by nuclear division without cytoplasmic division. Some umbrella cells can be smaller with degenerated (pyknotic) nuclei and mimic atypical cells.
Glandular cells can be columnar or cuboidal, generally seen in instrumented urine. They may be derived from cystitis cystica or glandularis or the female genital tract.
Renal tubular cells are small, generally poorly preserved with pyknotic, dark, eccentric nuclei and granular cytoplasm. They are easier to recognize when they form small clusters or casts.
Seminal vesicle cells have enlarged, hyperchromatic nuclei. High N/C ratio can also have prominent nucleoli, mimicking high-grade urothelial cells. Presence of yellow-brown cytoplasmic lipofuscin pigment and sperm in the urine specimen are important clues for identification. They can also cause abnormal DNA ploidy measurement.
Intermediate and basal cells often discriminate in clusters with stones, infection, or procedures. Cell clusters with urolithiasis can be 2-3D, spherical, with smooth cytoplasmic contours. Intermediate cell clusters may have a feathery appearance at the periphery. These clusters can be abundant in instrumented urine but can also be present in voided urine. When the clusters show no obvious cytological atypia and no fibrovascular core, they are considered “pseudo-papillae” and should be classified as NHGUC, but a comment could be added that LGUN cannot be ruled out without clinical correlation, as urine cytology has low sensitivity for detecting LGUN.
Polyomavirus: The typical findings have been described as “decoy cells,” “comet cells,” or “net cells” with large homogeneous opaque or ground glass intranuclear inclusions. However, in different stages of infection, the cytopathological features may not be typical. When viral particles leach out, chromatin may be coarse, mimicking HGUC; atypical diagnosis may be rendered.
BCG, mitomycin, and thiotepa may be intravesically administered. They cause inflammatory response, producing sloughing and degeneration of both benign and neoplastic urothelium. Mitomycin and thiotepa can cause nuclear enlargement, hyperchromasia, smudgy chromatin, and degenerated or vacuolated cytoplasm. Most have a low N/C ratio. When they have a high N/C ratio, they closely mimic HGUC. Systemically administered drugs like cyclophosphamide and busulfan may also cause marked cellular abnormalities. Radiation effects are characterized by cytomegaly with nuclear enlargement, multinucleation, and abundant vacuolated cytoplasm, maintaining a low N/C ratio.
Next is AUC: Atypical Urothelial Cells.
Diagnostic criteria are:
• Non-superficial, non-degenerated cells with increased N/C ratio >0.5
• Plus one of the following nuclear features:
› Nuclear hyperchromasia, OR
› Irregular, clumpy chromatin, OR
› Irregular nuclear membrane, OR
More pictures of AUC. They also have increased N/C ratio. Some clusters show irregular nuclear membrane. Some show nuclear hyperchromasia. Some show irregular, clumpy chromatin.
Next is HGUC: High-Grade Urothelial Carcinoma. Diagnostic criteria include:
• N/C ratio >0.7
• Moderate to severe nuclear hyperchromasia
• Irregular nuclear membrane
• Coarse, clumped chromatin
• Quantity also matters here; needs at least 5-10 abnormal cells for HGUC.
Other notable cytomorphological features include cellular pleomorphism, prominent nucleoli, mitoses, necrosis, etc. Urinary cytology has high sensitivity and specificity in detecting both high-grade papillary urothelial carcinoma and CIS.
Next is SHGUC: Suspicious for High-Grade Urothelial Carcinoma. Diagnostic criteria include:
• Increased N/C ratio and hyperchromasia
• Plus one of the following:
› Irregular, clumpy chromatin
› Irregular nuclear membranes
For N/C ratio, the TPS book uses 0.5-0.7, which is similar to AUC category. However, recent lectures from Dr. Wojcik and Dr. Barkan use N/C ratio >0.7, which is close to HGUC category. I hope TPS 2.0 can better clarify this discrepancy.
Quantity matters here. If only 27,000 cases for 2 years before and after TPS implementation.
Our data showed AUC decreased from 29% to 6.2%. Breaking down by specimen:
• Voided urine: AUC decreased from 27% to 6% (75% decrease)
• Instrumented urine: AUC decreased from 37% to 9% (75% decrease)
We also examined UroVysion FISH results and follow-up surgical results when available for different urinary cytology diagnoses.
• AUC associated with positive FISH increased from 16.7% to 37.6%.
• AUC associated with follow-up HGUC increased from 9% to 57%.
This table shows the cytology performance in detecting urothelial carcinoma. The performance of AUC for detecting HGUC was significantly improved (red font). Specificity increased from 49% to 86%. Positive predictive value improved from 9% to 39%. Accuracy improved from 50% to 85%.
In conclusion, implementing TPS resulted in a significant decrease in atypical diagnoses and significant improvement in specificity and PPV in detecting HGUC. AUC was significantly better correlated with UroVysion results, decreasing UroVysion test requests and saving medical costs. AUC should be considered a clinically relevant group requiring serious clinical workup in the TPS era.
Lastly, urinary cytology has low sensitivity in detecting LGUN. Sensitivity may be even lower in TPS. It is important for urologists to understand the limitations of urinary cytology and implications of changes introduced by TPS to best manage patient care.
Thank you."

[@essamahmed9242]

Thank you Dr Wei Tian for the excellent presentation and thanks to Shah for inviting Dr Wei Tian

[@mnagar]

Great presentation!! Thanks

[@aysesenamavus6036]

Thanks for the presentation

[@jdb853]

Lovely presentation, thank you