| Objective: Discussed in the condition of 1.5 T magnetic resonance imaging system using body coil, compare the PI- RADS v2 and Likert scale score method in the diagnosis performance of prostate cancer diagnosis.Methods: Collected 50 cases of postoperative pathology confirmed that prostate cancer patients with magnetic resonance imaging and clinical data from 2011- 2014. All cases were performed routine T2 WI and DWI, DCE. Check has not done before piercing, surgery or radiation and chemotherapy, etc. Separate the prostate images according to the peripheral and central area, glands, body, three levels, which is divided into 12 areas. successively with PI- RADS v2 and Likert scale to scores of all partitions, score results as compared with pathology results. The first part: the comparison of PI-RADS and Likert scale out(PZ) diagnosis efficiency. The second part: confirm cancerous area of high, medium and low into three teams, evaluate using PI- RADS v2 and Likert scale in three risk cancer detection sensitivity respectively.Results: mp- MRI images of the prostate in 50 cases were divided into 600 partition of the prostate(PZ: 300, CZ: 300), which achieve predefined(the biopsy results is clear prostate puncture, the puncture position to match the specific prostate partition) partition, 493, included in the study, invalid predefined(no pathological results or biopsy in the partition results cannot be determined in prostate specific partition) of 107 partitions, studies have been made. With the pathological results of effective predefined partition, a total of 210 cancer partition, the peripheral zone(PZ) : 165, the central gland(CZ) : 45, in cancerous partition, high-risk group of 46, moderate group 144, low-risk group 20; In a total of 283 effective predefined partitioning china-africa canceration partitions.Two doctors respectively according to the PI- RADS v2 grading standard, will be 1, 2, 3, 4, and 5 points for diagnosis of carcinoma and the value, as compared with the pathological results, diagnostic performance analysis. Results show that: the PI- RADS v2 score "4" is divided into diagnostic boundary value, the two physicians grading diagnosis: accuracy were 78.9%, 79.2%, sensitivity 57.3%, 70.4% respectively, specific degree 78.8%, 83.8% respectively, good consistency of diagnosis and pathology results(Kappa value 0.659 ~ 0.659).Two doctors respectively according to the criteria of Likert scale. Take 1, 2, 3, 4, and 5 points as threshold values, compared with the pathological results, diagnostic performance analysis. Results show that: the Likert scale score "4" for the diagnosis of boundary value, the two physicians grading diagnosis: accuracy were 79.8%, 80.7%, sensitivity 53.7%, 70.9% respectively, the specific degree 79.8%, 84.3% respectively, consistency of diagnosis and pathology results(Kappa value 0.579 ~ 0.579).Take "4" as the threshold value, compared with the pathological results, the central gland area: diagnostic accuracy Likert scale score(two physicians were 0.839, 0.839) were higher than PI- RADS v2 grading diagnosis accuracy(two physicians were 0.745, 0.745), between the difference was statistically significant(P < 0.05); Peripheral zone: two physicians PI- RADS v2 diagnostic accuracy rating as a Likert scale, the area under the ROC curve in 0.7 ~ 0.9, no differences between the two methods.Two doctors used PI- RADS v2 and Likert scale scoring criteria, respectively to obtain effective predefined grading diagnosis prostate partition, high scores are consistent with the diagnosis and pathology results(Kappa = 0.579 ~ 0.648), the difference between the two physician has no statistical significance(P > 0.05). Two doctors respectively using Likert scale score of the glands in consistency between diagnosis and pathology results(Kappa value are: 0.644, 0.629) were higher than in peripheral zone(Kappa value are: 0.579, 0.583), between the difference was statistically significant(P < 0.05).PI- RADS v2 and Likert scale was applied to high-risk group, moderate and low risk grading diagnosis of prostate cancer, the sensitivity of the detection of prostate cancer in the high-risk group were higher, the high-risk group of carcinoma(two doctors used PI – RADS v2 scoring diagnostic sensitivity of 97.3% and 95.6% respectively, Likert scale score diagnostic sensitivity of 85.6% and 86.9% respectively);High risk of cancer( PI- RADS v2 scores diagnostic sensitivity of 87.6% and 77.0% respectively, the sensitivity of the Likert scale score in the diagnosis of 81.6% and 75.5%, respectively). The prostate cancer in the high-risk group, the two doctors used PI- RADS v2 grading diagnosis sensitivity were higher than Likert scale, the difference between statistically significant(P < 0.05). In the low-risk group cancer(two doctors used PI- RADSv2 scoring diagnostic sensitivity of 75.5% and 63.7% respectively, the sensitivity of the Likert scale score in the diagnosis of 77.3% and 67.1%, respectively), PI – RADS v2 and Likert scale score diagnostic sensitivity than in high-risk groups, application of PI- RADS v2 and Likert scale to rate the diagnosis of the difference between the results of no statistical significance(P > 0.05).Conclusion: Both of PI- RADS v2 and Likert have good performance in the diagnosis of prostate cancer, and for the diagnosis of central gland area, Likert scale is better. In the high-risk group of prostate cancer, the sensitivity of using PI- RADS v2 is higher than the Likert scale. |
PDF Full Text Request