The ultrasound guided hollow needle biopsy is the so-called gold standard for the diagnosis of PC. In the US, according to 2002 cancer statistics, about 1,300,000 prostate biopsies were performed annually and detected 189,000 new cases of PC (15%) [82]. However, if the gold were "24 carat" for a single biopsy, it would never be necessary to repeat the procedure to "find" a suspected tumor or tumors. Unfortunately, this is not the case. The reason is that the procedure may miss the tumor or tumors because either they are small or in a part of the prostate that is not biopsied. Consider a small blueberry muffin containing only one or two blueberries and imagine trying to see if there are any at all with a few random thrusts of a hollow needle. Once a biopsy has identified the characteristic cells associated with PC, the diagnosis takes on a fairly high level of certainty. In fact, the nature, distribution and quantity of cells recovered that are suggestive or indicative of cancer are the basis of a scoring system, although the resultant scores are frequently changed (mostly upward) if a pathological study of the removed prostate is carried out. This is always done after a radical prostatectomy (RP), but those who elect radiation therapy do not have the opportunity for this definitive check on the information provided by their biopsy.

The importance of a correct diagnosis, including a correct assessment of the significance or seriousness of the disease if present, cannot be understated. The diagnosis of prostate cancer judged significant will most certainly lead to a discussion of treatment, and the two common treatments for localized cancers both have side effects that may profoundly alter an individual's quality of life. To have a radical prostatectomy or radiation therapy when in fact there is only insignificant or indolent cancer present is obviously something to avoid, although there are no doubt some who would debate this point. Likewise, to have a false negative diagnosis when an aggressive cancer is present is also highly undesirable unless the individual has a very short life expectancy. Thus the imperfections inherent in the standard diagnostic method are disturbing.

The prostate is relatively inaccessible and to acquire tissue sample requires an invasive procedure. Since sampling needles are used, the two most direct routes are through the wall of the rectum or through the perineum, which is the area between the anus and the scrotum. Almost all biopsies employ the former approach. The first reported needle biopsy occurred in 1930 and the first transrectal procedure was reported in 1937. The procedure was revolutionized in the mid 1980s with the advent of both a transrectal ultrasound device of adequate imaging resolution and its combination with a needle device which allowed the operator to direct needles under ultrasound guidance into the targeted sections of the prostate. Shortly thereafter a spring-loaded core biopsy device (a so-called gun) was developed that operated in conjunction with the ultrasound probe. Continued improvements in the resolution of the probe and the design of the needle gun have resulted in the present day instruments which provide both a satisfactory level of visualization coupled with the ability to direct the biopsy needle into a variety of regions of interest. The procedure can be done in an urologist's office, generally requires less than an hour, and is generally accompanied by minimal or at worst tolerable pain, especially if local anesthetic is used.

Some tumors can be imaged by the ultrasound device used in biopsy, and this can prompt targeting these areas for sampling. However, many tumors, especially early ones, are effectively invisible on the current grey-scale ultrasound [83]. Thus using ultrasound alone is not a solution to the diagnostic problem. While the ability to visualize and study the prostate gland has increased greatly with modern transrectal ultrasound, this has not resulted in substantially improved cancer detection without biopsy confirmation [83].

Since a biopsy is at present the only means of PC diagnosis, it is difficult to test its false negative rate since there is no "standard" for comparison. One can obviously not recruit a group of men who agree to radical prostatectomies after a negative biopsy to see if PC was missed! The solution to this dilemma is repeating the biopsy in cases where there is a high risk or a suspicious DRE, since repeating such an invasive procedure can be justified on medial grounds, and if enough biopsies are done, it turns out that a limiting and reasonably accurate estimate is achieved regarding the presence or absence of cancer. Men should be aware that repeat biopsies following a negative result in patients with high risk or suspicious DRE represent normal practice. The repeat procedures are done six to twelve weeks apart in order that the trauma associated with the previous test has resolved.

Biopsy Technique and Optimum Number of Cores

When ultrasound guided biopsies became routine, they initially involved taking six samples in the peripheral zone, the so-called sextant biopsy. In a sense this was a random biopsy since no attempt was made to target selected spots. It soon became apparent that the sextant procedure missed a significant number of cancers, the percentage frequently quoted being around 25%. Thus protocols that used a single sextant biopsy to determine the presence or absence of cancer in, for example, studies of the success of screening with PSA, were in fact using a rather imperfect diagnostic tool. That is, in general more subjects below a given PSA cut-off had PC than the biopsy data indicated. Also a single biopsy, when used to define a "cancer free" population for the purposes of a study, was in fact quite unsatisfactory. The situation is even more serious in subjects with large prostates where even a higher percentage of cancers are missed with the sextant protocol.

A recent study reported in 2005 provides an example of the problem of missed tumors. Djavan et al [84] studied 1051 men with risk defined by a PSA between 4 and 10 and compared the results of multiple biopsies with those obtained at the initial biopsy. A random sextant set plus two additional transition zone samples were used in all the biopsies. The first biopsy revealed an incidence of 22% PC. A second biopsy on those who were negative found 10% had PC giving an overall incidence for the two procedures of 30%. The third biopsy on the remaining "negatives" found 5% PC for an overall incidence of 33.3%. The fourth biopsy increased the incidence to only 33.6% overall. If one assumes that the fourth biopsy found most of the remaining cancer, the first biopsy missed 12% of the tumors present. A subset of patients had an RP. Organ-confined disease was present in 58% of cancer found in the first biopsy, 60.9% in the second, 86.3% in the third and 100% in the fourth. As might have been expected, cancers detected in the fourth biopsy had lower grade, stage and tumor volume as compared to those found in the first or second biopsy, i.e. the smaller the tumor or tumors, the harder they are to find. The third and fourth biopsies were accompanied by slightly higher level of morbidity than the first two. Thus an initial so-called extended biopsy protocol involving 8 cores still had a false negative rate of approximately 12%.

In a large serial biopsy study (2526 subjects), Roehl et al [85]found cancer detection rates of 29%, 17%, 14%, 11% 9% and 7% for one to six procedures. In this study the number of cores was variable but frequently the sextant scheme was used. In agreement with Djavan et al [84], there was a trend toward more organ confined cancers detected in subsequent biopsies as compared to the initial biopsy. Nearly 25% of the cancers ultimately detected were missed in the initial biopsy. Viewed another way, 77% of cancers were detected on first biopsy, 91% after two, 97% after three and 99% after four.

Both of these studies indicate the need for improving the first biopsy to decrease the number of repeats. Ochiai et al [86] have recently reviewed the current status of this problem. This review of the recent literature makes it clear that both altering and increasing the regions sampled and increasing the number of cores from 6 to 10 or 12 significantly increases the rate of cancer detection on first biopsy. This improvement is in part due to directing sampling to regions of the prostate with high probability of harboring tumors. For example, Presti et al [87] found that a more judicious use of even the sextant protocol yielded an 83% detection rate as compared to 78% with the random standard sextant approach, while a 10-core scheme which included high probability regions yielded 97% detection. The reference standard in this study was the 12-core biopsy which was not validated with repeat biopsies. Thus the number of missed cancers, while probably small, was in fact unknown, as it was also in the technique comparison studies reviewed by Ochiai. It in fact appears that the optimum number of cores and the core location strategy remains unknown if the ultimate concern is how many cancers were missed. Studies that compare two or more protocols employed during a single biopsy are relatively easy to accomplish since the protocols are combined in a single procedure, but to test several different protocols where three or four identical repeat biopsies are used to approximate the cancers missed in each protocol represents a much more complex study, both to organize and implement.

In what is called a saturation biopsy, a large number of cores, even over 40, are taken! General anesthetic is generally required and the procedure is accompanied by higher morbidity than observed in normal biopsy procedures. Scardino recommends against this protocol [2].

Current practice involves taking more cores than in the traditional sextant protocol, typically, 8, 10, 12 or 13. In addition, when a negative first biopsy is obtained with a patient judged at high risk because of a suspicious DRE, a very high PSA, or a family history of PC, one or more additional biopsies are generally recommended. After three or four negative results, the probability of cancer being present is low, typically less than 5%, but still finite and, unless the specific biopsy protocol has been studied locally with repeat biopsies, it is unknown. Coming back to the individual who wants to know for sure whether he has cancer, it appears from what is currently known that not only is a biopsy required, but one would not be enough unless a highly reproducible protocol was used that had been validated locally by repeat biopsies to establish the false negative rate and the initial biopsy was found to give a very low rate. One 12-core protocol yielding 2.3% false negatives on first biopsy has been reported, but the group studied was very small [64]. Thus journal articles can be misleading when a high detection rate is quoted for a particular protocol. For example, Philip et al [88] report that their 10-core biopsy technique found 98.6% of cancers in the group studied, but this was based on the reference standard of the number found with a 12 core protocol, not on repeat biopsies until there was a negligible increase in the discovery rate. It is also now fairly clear that men with large prostates require a larger number of cores to achieve the same detection rate as a man with a small prostate [86].

The ultrasound guided needle biopsy can also be used to investigate such questions as whether or not the cancer has spread to the surrounding fatty tissue, the seminal vesicles, or the prostate nerves. Information from such targeted sampling can be of great help in assessing the probable extent of the cancer. Thus a positive biopsy can sometimes prompt a repeat biopsy, the object being to obtain more information about the tumor or the presence of cancer in adjacent areas accessible in the transrectal biopsy. Such information may impact treatment decisions.

Biopsy Pain

The pain associated with the biopsy procedure arises mainly from the prostate itself rather than from the needle puncture of the wall of the rectum. In the past it has been generally considered that pain was not an issue. However, discomfort associated with the procedure is indeed experienced with 60-80% of men reporting mild to moderate pain and 6% of patients judging the pain so uncomfortable that they considered general anesthetic indicated [89]. In unusual cases, the pain may be so severe as to limit the number of cores that can be taken [90]. The current practice of taking up to 12 cores aggravates the problem [89]. Several approaches can be used to minimize this pain. One involves "conscious sedation" such as is used in colonoscopy. Complications from sedation are extremely rare. With sedation, one is awake, feels little pain, and generally does not recall the details of the procedure [37].

Another approach involves a local anesthetic gel (lidocaine gel) used to lubricate the ultrasonic probe, while a more direct approach involves injecting (infiltrating) lidocaine directly into the vicinity of the nerves leading into the prostate, a procedure done under ultrasound guidance. This latter technique is reminiscent of the dental nerve block used for work on the lower jaw. Studies that have either compared the use of lidocaine gel with infiltration or simply evaluated the infiltration method in a randomized fashion find the infiltration protocol superior to the gel and very effective in pain reduction [89-92]. However, lidocaine, even when used as a nerve block, is effective only for a few hours, and patients frequently report rebound pain by the evening of the biopsy. This can be relieved by the use of a long-lasting anti-inflammatory (diclofenac) administered via a suppository about an hour prior to the procedure [93]. The lidocaine nerve block became popular after the study of Soloway and Obek was published in 2000 [94].

If pain can be controlled, what else is there to worry about in connection with having a prostate biopsy? Two things come to mind. One is dealing with positive results that require a decision regarding treatment or watchful waiting but also may induce psychological stress, perhaps severe, associated with the knowledge that cancer is present. The other involves the potential for complications associated with the needle biopsy. The incidence of biopsy-related complications has received considerable study. Complications include blood in the urine and semen (hematuria and hematospermia), prostate or urinary infection, a serious or even life-threatening systemic infection, an acute urinary retention episode, temporary erectile dysfunction, and finally rectal bleeding, which if severe can require immediate intervention including surgery. Also, moderate to severe so-called vasovagal responses can occur with drops in systolic blood pressure requiring intravenous fluid administration and in severe cases, with neurological events such as seizure or loss of consciousness. The complications of the prostate biopsy merit further discussion.

Biopsy Complications

Since the rectum is not sterile there is risk of the needle puncture carrying intestinal bacterial into the prostate with the resultant possibility of infection. To guard against this eventuality, it is now standard practice to prescribe antibiotics both before and after the procedure, with the fluoroquinolone class the most popular (e.g. ciprofloxacin, levofloxacin or ofloxacin). However, even with antibiotic prophylaxis, there is still a small risk of post-biopsy infection. In rare instances the fluoroquinolone class of antibiotic can cause very serious side effects.

It is difficult to quote meaningful combined statistics from studies of complications due to a variation in definitions, the failure to distinguish early from delayed morbidity, mild from severe complications and the lack of stratification with regard to antibiotic use. Reviews of large numbers of studies produce huge ranges in percent incidence. For example, Ghani et al [95] give the following ranges:

• Hospitalization - 0-4%
• Hematuria - 12.5-80%
• Hematospermia - 1.3-58%
• Urinary retention - 0.2-10%

It is perhaps more meaningful to look at a study where early and delayed as well as mild vs. severe complications were distinguished. Djavan et al [96] examined the complication rate associated with 1051 initial biopsies which were all done with fluoroquinolone antibiotic prophylaxis preoperatively and for 4 days thereafter. For early morbidity, between 97.2 and 98.2% of patients had no or mild complications with the exception of mild hematuria, which was seen in 38% of cases. For moderate to severe early complications, they found rectal bleeding 2.1%, hematuria 62%, vasovagal episodes 2.8%, and urinary retention 1.8%. Delayed morbidity was as follows:

• Urinary tract infections -10.9%
• Fever - 2.9%
• Urinary tract infections with fever - 2.1%
• Urinary retention - 0.9%
• Hematospermia - 9.8%
• Recurrent mild hematuria -15.9%
• Persistent dysuria (difficulty or pain during urination) - 7.2%

The authors also present a review of eleven studies. In all but two cases antibiotic prophylaxis was used. The range of incidence was smaller than reported by Ghani et al [95] but still rather large.

Rectal bleeding can be serious. Scardino [2] gives a rate of 1 in 1000 patients that required hospitalization for transfusion, cauterization or sutures to resolve biopsy induced severe rectal bleeding. He also quotes a rate of about 3% for urinary tract infections or prostatitis if antibiotic prophylaxis is used, with 6-10% otherwise. He makes a very strong point as does Walsh [37] that patients must be very vigilant after a biopsy for symptoms of complications that can quickly become severe. The warning signs are a fever of 101°F (38.3° C) or higher, chills, muscle aches, or urinary problems such as abnormal frequency, urgency or burning. In his opinion, immediate action is indicated, at an emergency room if necessary, in order that the infection apparently present can be cultured and treated. Patients should inform the ER staff or attending physician of the recent transrectal biopsy. Failure to take immediate action (even a 6-12 hour delay is serious) can in some cases result in an overwhelming septic condition with attendant long hospitalization, serious health risks and even death. Scardino gives a risk of 1 in 200 of this scenario. Because of the risk of delayed infection or bleeding, Scardino suggests avoiding long air trips or travel to remote areas where immediate medical attention is unavailable until the risk of these complications is very low.

Biopsy candidates who have previously undergone total joint replacements are at increased risk of hematogenous prosthetic joint infections. The American Academy of Orthopedic Surgeons (AAOS) and the American Urological Association (AUA) have issued a joint advisory statement regarding this potential problem. See http://aaos.org/wordhtml/papers/advistmt/1023.htm. Those at increased risk include all patients during the first two years after joint replacement, patients immuno- compromised or immuno-suppressed, and patients with previous prosthetic joint infections, malnourishment, hemophilia, HIV infections, diabetes and malignancy. Patients corresponding to one or more of these criteria should inform their urologist of the situation prior to a biopsy. The AAOS/AUA recommendations provide a suggested special antibiotic prophylaxis regimen.

It is important to realize that most of the complications of the transrectal ultrasound guided biopsy are mild and short lived. The use of anesthetic, especially the prostatic nerve block, should reduce the procedural pain to a minimum, and while severe complications are unpleasant to contemplate and of finite but low probability, it is hard to see that they should be a deterrent to having a medically justified biopsy. However, as has been discussed at length above, what is medically justified is sometimes far from clear and subject to debate.

The biopsy generates a set of tissue samples (cores) approximately 0.4 mm in diameter and 15 mm in length which are then examined by a pathologist who will look for cells characteristic of PC and characterize the type and extent in each core. The number of cores yielding positive results is reported and in most studies is correlated with the risk of non-organ confined disease. Finally, a score, called the Gleason Score is calculated and forms part of the basis of clinical judgments as regards treatment and prognosis. While this is the now a widely accepted protocol for the analysis of biopsy cores, it turns out to be somewhat less than perfect in either identifying the seriousness of the cancer or the prognosis. It is, however, all that there is available aside from PSA levels and DRE observations to guide the urologist down the path to treatment or watchful waiting unless the patient presents with symptoms of advanced disease or distant metastasis. These scores are also used to judge prognosis. Pathologists also grade the cancer with the Gleason system when it is found in tissue samples recovered during the TURP surgical procedure for BPH. Gleason Scores can also be calculated from an examination of the prostate after an RP.

Part III of this report will discuss the significance of the Gleason score, the staging of the cancer and the ultimate prognosis.