Purpose <p>To determine whether anatomical location and apparent diffusion coefficient (ADC) risk stratification optimizes biopsy selection for PI-RADS 3 lesions in the transition zone (TZ).</p> Methods <p>This retrospective single-center study included 218 TZ PI-RADS 3 lesions in 128 patients who underwent radical prostatectomy between January 2014 and September 2023. Two radiologists identified 218 TZ PI-RADS 3 lesions in 128 patients on MRI and measured mean ADC; a third radiologist correlated imaging findings with whole-mount histopathology. Lesions located in the anterior apex, anterior mid-gland, or middle apex were classified as Location-risk positive; lesions with mean ADC &lt; 0.800 × 10⁻³ mm²/s were classified as ADC-risk positive. Four virtual biopsy strategies were compared (conventional, Location-based, ADC-based, and Location + ADC-based). Outcomes included cancer detection rate (CDR), biopsy burden (performed/avoided), clinically significant prostate cancer (csPCa) yield (detected/missed), positive predictive value (PPV), and net benefit on decision curve analysis (thresholds, 5–25%).</p> Results <p>Mean age was 68 ± 5.9 years (median, 69; range, 50–78). Among 218 lesions, 34 (16%) were csPCa. CDR was higher in location-risk positive than negative lesions (43% [28/65] vs. 4% [6/153]; <i>p</i> &lt; 0.001) and in ADC-risk positive than ADC-risk negative lesions (47% [15/32] vs. 10% [19/186]; <i>p</i> &lt; 0.001). Compared with conventional biopsy, the Location + ADC-based strategy reduced biopsies by 62% while detecting 31 of 34 csPCa lesions (91%) and missing 3 of 34 csPCa lesions (9%). Both location-informed strategies showed higher net benefit than the conventional across the evaluated thresholds.</p> Conclusion <p>For TZ PI-RADS 3 lesions, combining anatomical location with an ADC threshold reduces unnecessary biopsies while preserving csPCa detection.</p>

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

Location- and ADC-based stratification of transition zone PI-RADS 3 lesions to optimize biopsy indication

  • Yasuhiro Enjoji,
  • Hiromi Edo,
  • Hiroshi Kuwamura,
  • Taiki Nozaki,
  • Hiroaki Kobayashi,
  • Keiichi Ito,
  • Kosuke Miyai,
  • Susumu Matsukuma,
  • Hiroshi Shinmoto

摘要

Purpose

To determine whether anatomical location and apparent diffusion coefficient (ADC) risk stratification optimizes biopsy selection for PI-RADS 3 lesions in the transition zone (TZ).

Methods

This retrospective single-center study included 218 TZ PI-RADS 3 lesions in 128 patients who underwent radical prostatectomy between January 2014 and September 2023. Two radiologists identified 218 TZ PI-RADS 3 lesions in 128 patients on MRI and measured mean ADC; a third radiologist correlated imaging findings with whole-mount histopathology. Lesions located in the anterior apex, anterior mid-gland, or middle apex were classified as Location-risk positive; lesions with mean ADC < 0.800 × 10⁻³ mm²/s were classified as ADC-risk positive. Four virtual biopsy strategies were compared (conventional, Location-based, ADC-based, and Location + ADC-based). Outcomes included cancer detection rate (CDR), biopsy burden (performed/avoided), clinically significant prostate cancer (csPCa) yield (detected/missed), positive predictive value (PPV), and net benefit on decision curve analysis (thresholds, 5–25%).

Results

Mean age was 68 ± 5.9 years (median, 69; range, 50–78). Among 218 lesions, 34 (16%) were csPCa. CDR was higher in location-risk positive than negative lesions (43% [28/65] vs. 4% [6/153]; p < 0.001) and in ADC-risk positive than ADC-risk negative lesions (47% [15/32] vs. 10% [19/186]; p < 0.001). Compared with conventional biopsy, the Location + ADC-based strategy reduced biopsies by 62% while detecting 31 of 34 csPCa lesions (91%) and missing 3 of 34 csPCa lesions (9%). Both location-informed strategies showed higher net benefit than the conventional across the evaluated thresholds.

Conclusion

For TZ PI-RADS 3 lesions, combining anatomical location with an ADC threshold reduces unnecessary biopsies while preserving csPCa detection.