Objectives: The most common presentation of knee osteochondritis dissecans (OCD) is a stable lesion on the lateral aspect of the medial femoral condyle (MFC) in an adolescent or pre-adolescent athlete. The standard of care for primary treatment is non-operative, and includes rest/activity modification and often weight bearing protection or bracing. Failed conservative management often leads arthroscopy and drilling of the lesion. Two different primary drilling techniques have been utilized, but no prospective studies have compared their relative effectiveness. The study hypothesis was that retro-articular drilling (RAD), the slightly newer technique, would not be inferior to trans-articular (TAD), with regard to rate of healing, time to return to sports (RTS), and patient-reported outcome scores (PROs). Methods: Skeletally immature (n=113) patients presenting with MRI-confirmed stable OCD of the MFC who did not demonstrate substantial healing after a minimum of 3 months of non-operative treatment were prospectively enrolled by one of seventeen surgeon-investigators (at 14 centers, representing all major geographic regions in the U.S.) and randomized to TAD or RAD. Post-operatively, serial radiographs were obtained every 6 weeks to assess healing, and PROs were obtained at 6 months, 12 months, and 24 months. Twelve patients were closed out at time of surgery due to lesion instability detected during arthroscopy. Power analysis determined that in order to detect a difference in 2-year IKDC score between RA and TA groups with 80% power, sample sizes of 37 subjects per group would be required if the true standard deviation were 15. This analysis was based on conducting an independent samples Student’s t-test with alpha set to 5%. Results: Ninety-one study subjects were included, consisting of 51 TAD and 40 RAD patients, respectively, with the two groups being similar in age (12.6 years vs. 11.9 years), sex distribution (45% vs. 27% female, p=0.081), and 2-year PRO response rate (both 90%). No significant differences between TAD and RAD were detected in follow-up Pedi-IKDC, Lysholm, Marx knee activity score, or KOOS QOL scores (Table 1). Revision/additional OCD surgery occurred in 10% of patients in RAD and 4% in TAD (p=0.40). 73% of TAD patients reached a ‘healed’ status at a mean of 1.15 years, compared with 60% RAD patients at a mean of 1.21 years. Conclusions: While both primary forms of OCD drilling (TAD and RAD) showed consistent post-operative healing, achieving a completely ‘healed’ status was often a more prolonged process, taking approximately 1 year, despite clinical improvement and RTS being achieved much sooner. PROs were similar between drilling techniques. Significantly higher powered studies are needed to better elucidate the greater revision surgery rates in RAD compared with TAD, but overall risk is low and absolute risk only 6%. The current data support either drilling technique, which may be technically simpler, without the need for fluoroscopy, with TAD, and may be more protective of the chondral articular surface with RAD.