• Slight changes in inclusion criteria to avoid contradictions to exclusion criteria; • Details on consequences of patients requiring subsequent surgery were added (patients with subsequent surgery were allowed to remain in the study); • Specifications on confirmatory hypothesis tests of secondary efficacy endpoints were added; • An additional MRI assessment was scheduled for the screening visits (to assess eligibility criteria).

• Participating countries were updated (France, Poland, Netherlands, Czech Republic); • Inclusion criterion No. 2 was generalized from "patient has symptomatic knee pain indicative of articular cartilage defects of the knee unsuccessfully treated with conservative care (e.g., analgesics, rest, and physical therapy)" to the current wording: "Patient has a localized articular cartilage defect of the femoral condyle or the trochlea of the knee. 2 localized cartilage defects are accepted if the total defect size is ≤ 6 cm2, both cartilage defects are located at the femoral condyle and/or the trochlea and both cartilage defects are to be treated with NOVOCART® 3D plus or microfracture."; • The allowed defect size was extended from "≥3 and ≤5 cm2" to "≥2 and ≤6 cm2 post debridement"; • The maximum allowed meniscus resection was extended from 33% to 50%; • The exclusion criterion "patient has had prior release and excision of scar tissue except isolated lateral release in the target knee." was deleted, as this criterion would have created an artificial situation; • A couple of exclusion criteria were deleted, as their clinical relevance and impact on study results was not evident. However, the exclusion criterion concerning prior surgical treatments of clinical relevance in the target knee was added instead; • The excluded use of steroids within the past 3 months prior to screening was limited to systemic or intraarticular steroids within 30 days prior to tissue harvest/MFx; • Low-dose treatment with NSAIDs (not capable of pain relief) was allowed; • The accepted BMI was increased from 30 kg/m2 to 35 kg/m2;

• Based on supporting literature data showing the full evolvement of the beneficial treatment effects of MACT and MFx already after 24 months post-surgery, it was decided to prepone the time point for the primary efficacy analyses from 36 months to 24 months post-treatment. • A definition for treatment failure was included and the proportion of treatment failures was added as a secondary endpoint; • The originally planned patient number for the MRI sub-study (64 in each arm) was adapted to match the actual randomization ratio of 2:1 • Biomarker sampling (blood and urine) was limited from all patients to a subset of patients from pre-selected sites; • The screening period (which was previously unlimited in time) was restricted to a maximum of 3 months from screening visit 1 to visit 2 (arthroscopy); • Clarifications and slight changes in in- and exclusion criteria, e.g., clarification that, if 2 localized defects have to be treated, the size of each individual lesion had to be ≥2 cm2; permission to include patients with known history of diabetes, primary hyperparathyroidism or hyperthyroidism, provided these conditions were sufficiently controlled; the exclusion criterion "known history of cancer" was limited to the past 5 years; in the original protocol, patients taking indomethacin or other NSAIDs were not to be included into the study. This exclusion criterion was deleted and the concomitant medication section was revised accordingly to cover washout periods; • For unscheduled visits, the IKDC objective physician score, physical examination and vital signs, SF-36, KOOS, IKDC subjective score, activity level and functional status were added as mandatory assessments; • The post-treatment rehabilitation regimen according to Hirschmüller et al. was changed from a mandatory to an optional (although still recommended) procedure. Still, patients were to undergo post-treatment rehabilitation according to the standard of care in the respective site/country at least

• In the context of the "Pediatric Investigational Plan" (PIP) discussions, the Pediatric Committee (PDCO) requested prospective data on the efficacy and safety of NOVOCART® 3D in skeletally mature pediatric patients. As suggested by the PDCO, the protocol was amended to include pediatric patients with closed epiphyseal growth plates. The inclusion criteria were expanded accordingly and the statistical section updated to include the subgroup analyses; • The definition of treatment failures was amended to make clear that only graft/MFx-related conditions requiring surgical re-intervention should be considered treatment failures (in the previous definition the relatedness was missing); • Latvia, Lithuania, Hungary, and Turkey were added as participating countries to the study; • HTLV-I antibody testing was introduced for all patients living in, or originating from, high-prevalence areas, or with sexual partners originating from those areas, or where the patient’s parents originated from those areas; • Cryopreservation was explicitly allowed as an exception in the event that the originally agreed transplantation date had to be postponed; • It was added that patients in the MFx arm receiving other subsequent treatments on the target lesion (e.g., other ACT, total knee replacement etc.) were to be withdrawn from the study, whereas patients in the N3D arm were to remain in the study (but only adverse events related to NOVOCART® 3D plus treatment and serious adverse events (irrespective of relationship) were to be collected in these patients after change in therapy);

• The group of "further exploratory endpoints" (i.e., KOOS total responder rate, IKDC responder rate, rate of surgical re-intervention on the target knee) was added to the set of efficacy analyses; • A second definition of treatment failure (definition No. 2) was introduced to the effect that surgical re-interventions not affecting the closed surface of the transplant / microfracture area were not automatically classified as treatment failure; • Advanced T2 mapping analyses using GLCM features were added to the MRI analyses, since these features serve as a robust quantitative marker for collagen fiber organization; • The length of incision was omitted as key secondary endpoint (because this variable is recorded only in the N3D group); • Following an EMA recommendation, the IUDR procedure was used as replacement strategy for missing values in the primary and key secondary analyses. The linear mixed effect model for repeated measurements (MMRM) without imputation of missing values was implemented as a sensitivity analysis. The imputation of missing values due to treatment failure by "last observation carried forward (LOCF)" was omitted; • The nominal significance levels at the interim and main (final) analysis were fixed in order to ensure independency from the actual information available after the interim analysis (as recommended by the EMA); • A supplementary NI analysis was implemented in the case of a failed superiority test of NOVOCART® 3D plus vs. microfracture.