Here are a couple of quotes from an article that can be found at this link: http://www.medscape.com/viewarticle/481628
(It requires registration but it is free).
Voided urine cytology is far more specific for bladder cancer than urinalysis, but it lacks adequate sensitivity, particularly when used to detect the presence of low-grade tumors. In a systematic review comparing cytology with a number of bladder tumor markers (including NMP-22) (Lotan & Roehrborn, 2003), the specificity of urine cytology was 99%. Unfortunately, its sensitivity was far lower, varying from 30% to 50% in invasive bladder tumors and 20% to 30% in superficial malignancies. The relatively low sensitivity of cytology makes it inadequate as a screening modality. However, urine cytology is often used in combination with urethrocystoscopy to detect bladder cancers in persons deemed to be at high risk, particularly when urinalysis reveals clinically relevant hematuria.
While the preponderance of evidence reveals that NMP-22 is more sensitive than either urinalysis or voided urine cytology for detecting low-stage bladder cancers, it lacks adequate sensitivity to replace urethrocystoscopy for diagnosing bladder cancer (see Table 2 ). Data from a meta-analysis comparing voided urine cytology and NMP-22 supports the use of this (or the BTA STAT assay) over urine cytology because of its increased sensitivity in detecting low-stage and low-grade recurrences, but the authors also note that its role is limited to adjunctive to urethrocystoscopy. Until the sensitivity of the NMP-22 is improved, or another urine assay emerges that provides significantly greater sensitivity while maintaining adequate specificity, urethrocystoscopy will remain the reference test for detecting and diagnosing bladder and its recurrence.
With that said, when I had my recurrence (two years after initial diagnosis and treatment) it was the NMP22 test that showed something was there when the doctor couldn't see anything with the scope. So, it seems like having both done on a regular basis after initial diagnosis (every three months) is prudent and is considered the standard of care in the NCCN bladder cancer treatment guidelines.
My urine tests never showed cancer cells, even when I had 5 tumors. I even took them samples that were bright red and had large blood clots. When I finally got referred to a uro that knew what was going on he didn't even want a sample, he went straight for the scope. He said if I would have went any longer it would have been a lot worst.
I had urine cytology (looking for cancer cells in the urine) done when I first had bleeding, as part of the diagnostic workup looking for bladder cancer or other possible cause of the bleeding; they also checked then to see if there was a bladder infection. However, I was told even at that time that cytology rarely shows anything in the case of low grade tumors. Because my initial tumor, and later recurrence, were low grade non-invasive tumors, they've never done cytology again, as cystoscopic exam is the best way to keep on top of things.
There is another type of diagnostic test, which I know numbers of people here get, called FISH for short; I believe it checks for specific celluar DNA changes associated with bladder cancer. This is supposed to be more accurate than "conventional" cytology.
Small TA Grade 1, May-06; recur (2 tiny), same, June-08; TURBTs both times. BCG begun July-08, dosage to 1/3rd May-10, completed treatment December-11. All clear since 2008.