{
  "analysis_disposition": "amended_or_exploratory",
  "audit_time_label_sensitivity": {
    "all_sensitivity_gates_passed": true,
    "candidate_changed": false,
    "claim_boundary": {
      "automatic_annotation_change_permitted": false,
      "clinical_utility_proven": false,
      "independent_confirmation": false,
      "natural_error_detection_proven": false,
      "pathologist_error_detection_proven": false
    },
    "disposition": "post_confirmation_exploratory_sensitivity_only",
    "fold_assignment_label_source": "observed_label",
    "source": {
      "path": "artifacts/puma_audit_time_label_sensitivity/results.json",
      "sha256": "8f524b236995a495048a0955ebf930e14e732a1214d3856d3711571af13fd5cd",
      "size_bytes": 9944
    },
    "study_id": "puma_audit_time_label_sensitivity_v1"
  },
  "automatic_annotation_changes": false,
  "confirmatory_completed": false,
  "controlled_external_benchmark": {
    "all_success_conditions_met": false,
    "analysis_disposition": "prospectively_frozen_controlled_external_benchmark",
    "claim_boundary": {
      "clinical_or_operational_utility_proven": false,
      "natural_pathology_error_detection_proven": false,
      "pathologist_error_proven": false,
      "source_annotations_modified": false
    },
    "dataset": "MoNuSAC2020",
    "decision": "not_supported",
    "development": {
      "controlled_corruption_count": 2961,
      "controlled_corruption_rate": 0.1,
      "eligible_nuclei": 29610,
      "patient_group_count": 44
    },
    "primary_downstream": {
      "candidate_macro_f1": 0.5093608506212538,
      "corrupted_uncorrected_macro_f1": 0.5038352361344909,
      "minus_corrupted_uncorrected_interval_95": [
        -0.001505873468356683,
        0.012832750726675505
      ],
      "minus_corrupted_uncorrected_macro_f1": 0.00552561448676292,
      "minus_mean_matched_random_interval_95": [
        -0.0086920112069291,
        0.008485812521403327
      ],
      "minus_mean_matched_random_macro_f1": 3.0946417352573086e-05
    },
    "primary_ranking": {
      "changes_found": 1035,
      "difference_interval_95": [
        0.0991810700012709,
        0.18849133691115186
      ],
      "mean_matched_random_precision": 0.5560094530722484,
      "precision": 0.6988521269412559,
      "reviewed_count": 1481
    },
    "source": {
      "path": "artifacts/monusac_external_validation/results.json",
      "sha256": "b2724e3e0baedcd0f1eb0fc7dfae127bf3789b03ac626d2701477ff4bae8e7d4",
      "size_bytes": 33249
    },
    "study_id": "monusac_current_aanca_controlled_external_v1",
    "success_conditions": {
      "important_class_recall_non_degradation": false,
      "primary_downstream_beats_corrupted_uncorrected": false,
      "primary_downstream_beats_mean_matched_random": false,
      "primary_top_k_beats_exact_matched_random": true
    },
    "test": {
      "eligible_nuclei": 15494,
      "patient_group_count": 25
    }
  },
  "diagnostic_system": false,
  "external_validation": {
    "claim_boundary": {
      "biological_truth_proven": false,
      "clinical_utility_proven": false,
      "pathologist_error_proven": false
    },
    "completion_stage": "EXTERNAL_VALIDATION_COMPLETE",
    "overall_conclusion": "not_supported",
    "primary_subset": {
      "ap_minus_prevalence_interval_95": [
        0.006105441307506995,
        0.21362402839313893
      ],
      "average_precision": 0.07348905384277785,
      "downstream_success_conditions_met": false,
      "guided_minus_uncorrected_macro_f1": -0.014633352536513322,
      "guided_minus_uncorrected_macro_f1_interval_95": [
        -0.026683314580239315,
        -0.002414596650361811
      ],
      "natural_disagreement_count": 27,
      "natural_disagreement_prevalence": 0.03329223181257707,
      "patient_group_count": 5,
      "precision_at_5_percent": 0.0975609756097561,
      "precision_at_5_percent_minus_prevalence_interval_95": [
        -0.03007518796992481,
        0.15407470288624786
      ],
      "ranking_success_conditions_met": false,
      "sample_count": 811,
      "subset": "unbiased_control"
    },
    "reference_interpretation": "natural NP-label disagreement relative to inferred pathologist consensus; not guaranteed biological truth",
    "source": {
      "path": "artifacts/nucls_external_validation/unbiased-v1/results.json",
      "sha256": "a931100a57e8d4b2a34a0216f047de8aa9d21c7275bc1beafd3b361fd955e9f6",
      "size_bytes": 19653
    },
    "study_id": "nucls_natural_label_external_validation_v1"
  },
  "external_validation_completed": true,
  "natural_data_action": {
    "action": "retain_uncorrected",
    "reason": "official corrected and uncorrected single-rater releases are distinct FOV quality tiers and the raw database preserves no paired pre/post class label per nucleus",
    "source": {
      "path": "artifacts/nucls_supervised_qc_feasibility/results.json",
      "sha256": "e95dd1416ad2be08126aa710a4d00ff999a4cb25936fcf1f9fc1f71f62d36229",
      "size_bytes": 2949
    },
    "status": "unavailable",
    "study_id": "nucls_supervised_qc_prospective_v1"
  },
  "new_source_confirmation": {
    "all_success_conditions_met": true,
    "analysis_disposition": "prospectively_frozen_new_source_controlled_confirmation",
    "claim_boundary": {
      "automatic_annotation_change_permitted": false,
      "clinical_utility_proven": false,
      "controlled_noise_transfer_if_positive": true,
      "natural_error_detection_proven": false,
      "pathologist_error_detection_proven": false
    },
    "dataset": "PUMA",
    "decision": "controlled_noise_transfer_supported",
    "downstream": {
      "candidate_macro_f1": 0.6463102793504962,
      "mean_matched_random_macro_f1": 0.6382432313302706,
      "minus_matched_random": 0.008067048020225648,
      "minus_matched_random_interval_95": [
        0.0040931053452727805,
        0.011946725717818258
      ],
      "minus_uncorrected": 0.006426160085933802,
      "minus_uncorrected_interval_95": [
        0.003657212432103269,
        0.009365455935978527
      ],
      "uncorrected_macro_f1": 0.6398841192645623
    },
    "final_case_groups": 62,
    "final_nuclei": 30397,
    "retrieval": {
      "candidate_precision": 0.5377386634844868,
      "difference": 0.3233591885441528,
      "interval_95": [
        0.2592505925988822,
        0.3849444213529693
      ],
      "mean_matched_random_precision": 0.2143794749403341
    },
    "sources": {
      "results": {
        "path": "artifacts/puma_new_data_confirmation/results.json",
        "sha256": "e547d77a2ae5bffdaed9f64f83c369292fe11cd2e8a305d02fc3d2271131dd92",
        "size_bytes": 42958
      },
      "verification": {
        "path": "artifacts/puma_new_data_confirmation/verification.json",
        "sha256": "2f4b6d6c64d8dfa493a89be0111ca8752257dabce808d4d9fa35933cdd9c5d68",
        "size_bytes": 1268
      }
    },
    "study_id": "puma_new_data_confirmation_v1",
    "success_conditions": {
      "all_four_seed_directions_positive_against_both_controls": true,
      "all_hash_group_split_and_final_fold_guards_passed": true,
      "all_required_models_converged": true,
      "downstream_macro_f1_lower_bound_gt_matched_random": true,
      "downstream_macro_f1_lower_bound_gt_unchanged": true,
      "every_primary_class_recall_lower_bound_gte_minus_0_01": true,
      "retrieval_precision_lower_bound_gt_matched_random": true
    },
    "verification": {
      "all_44_models_converged": true,
      "all_seven_frozen_success_gates_passed": true,
      "source_reference_labels_unchanged": true,
      "verified": true
    }
  },
  "next_phase": {
    "natural_auto_change_allowed": false,
    "promotion_rule": "promote only if ranking, downstream confidence intervals, class safety and prospective workflow utility all pass",
    "required_gates": [
      "blinded multi-rater natural-case reference with ambiguity and abstention",
      "nested group-safe measured-utility queue development",
      "one policy frozen before any new confirmation outcomes",
      "one-shot untouched patient or WSI external confirmation",
      "prospective multi-site AANCA-versus-control workflow evaluation"
    ],
    "stage": "INITIALISED",
    "working_name": "AANCA v2 research phase"
  },
  "pannuke_qc": {
    "cross_class_overlap_patch_count": 575,
    "cross_class_overlap_pixel_count": 4318,
    "fold_count": 3,
    "no_class_arbitration": true,
    "overlap_touching_instance_count": 1411,
    "overlay_sha256": "a1bd87dd397417d711d1d4937429eae5f5d972d3fa6ffa27a45129339587f10a",
    "patch_count": 7901,
    "selection_sha256": "09886588591d9ebb9a725db1022bb0ab8fb94b4bcca419b486e2549b0cc5fd36",
    "source_masks_modified": false,
    "void_patch_count": 162,
    "void_pixel_count": 10486091
  },
  "presentation_status": "DEMO_COMPLETE",
  "primary": {
    "analysis_disposition": "amended_or_exploratory",
    "artifact_count": 2288,
    "artifact_manifest_sha256": "9abff1b2f0e745a50b3aa1922d3d725bb629276f6090c32e9b423fea82d0e0ce",
    "artifact_root_sha256": "8c1c7b277d96889dc4fb45aee282e77e3d351f687990e03e6b57ec5f2313c7e4",
    "bootstrap_iterations_required": 2000,
    "comparison_count": 36,
    "comparisons": [
      {
        "comparison_id": "h1_self_confidence_minus_random_confusion_seed_404",
        "interval_95": [
          0.2709081859665893,
          0.29664458251449377
        ],
        "method_a": "self_confidence",
        "method_b": "random_review",
        "metric": "average_precision",
        "p_value_holm": 0.005997001499250375,
        "point_difference": 0.28438693143936744,
        "status": "reported",
        "valid_bootstrap_iterations": 2000
      },
      {
        "comparison_id": "h1_self_confidence_minus_random_confusion_seed_405",
        "interval_95": [
          0.27864192876536475,
          0.3028247204060658
        ],
        "method_a": "self_confidence",
        "method_b": "random_review",
        "metric": "average_precision",
        "p_value_holm": 0.005997001499250375,
        "point_difference": 0.2909951271603709,
        "status": "reported",
        "valid_bootstrap_iterations": 2000
      },
      {
        "comparison_id": "h1_self_confidence_minus_random_confusion_seed_406",
        "interval_95": [
          0.27779687541231035,
          0.3031654200542621
        ],
        "method_a": "self_confidence",
        "method_b": "random_review",
        "metric": "average_precision",
        "p_value_holm": 0.005997001499250375,
        "point_difference": 0.2907067931408497,
        "status": "reported",
        "valid_bootstrap_iterations": 2000
      },
      {
        "comparison_id": "h1_self_confidence_minus_random_group_conditional_seed_404",
        "interval_95": [
          0.28890514216025037,
          0.3149602238192935
        ],
        "method_a": "self_confidence",
        "method_b": "random_review",
        "metric": "average_precision",
        "p_value_holm": 0.005997001499250375,
        "point_difference": 0.30188350175581824,
        "status": "reported",
        "valid_bootstrap_iterations": 2000
      },
      {
        "comparison_id": "h1_self_confidence_minus_random_group_conditional_seed_405",
        "interval_95": [
          0.28304403233500064,
          0.30838858657288076
        ],
        "method_a": "self_confidence",
        "method_b": "random_review",
        "metric": "average_precision",
        "p_value_holm": 0.005997001499250375,
        "point_difference": 0.2956436304787686,
        "status": "reported",
        "valid_bootstrap_iterations": 2000
      },
      {
        "comparison_id": "h1_self_confidence_minus_random_group_conditional_seed_406",
        "interval_95": [
          0.28028729899840366,
          0.3059159473434755
        ],
        "method_a": "self_confidence",
        "method_b": "random_review",
        "metric": "average_precision",
        "p_value_holm": 0.005997001499250375,
        "point_difference": 0.2931289952332816,
        "status": "reported",
        "valid_bootstrap_iterations": 2000
      },
      {
        "comparison_id": "h1_self_confidence_minus_random_instance_dependent_seed_404",
        "interval_95": [
          0.031088090225755212,
          0.041555304340355384
        ],
        "method_a": "self_confidence",
        "method_b": "random_review",
        "metric": "average_precision",
        "p_value_holm": 0.005997001499250375,
        "point_difference": 0.03620591936690898,
        "status": "reported",
        "valid_bootstrap_iterations": 2000
      },
      {
        "comparison_id": "h1_self_confidence_minus_random_instance_dependent_seed_405",
        "interval_95": [
          0.031088090225755212,
          0.041555304340355384
        ],
        "method_a": "self_confidence",
        "method_b": "random_review",
        "metric": "average_precision",
        "p_value_holm": 0.005997001499250375,
        "point_difference": 0.03620591936690898,
        "status": "reported",
        "valid_bootstrap_iterations": 2000
      },
      {
        "comparison_id": "h1_self_confidence_minus_random_instance_dependent_seed_406",
        "interval_95": [
          0.031088090225755212,
          0.041555304340355384
        ],
        "method_a": "self_confidence",
        "method_b": "random_review",
        "metric": "average_precision",
        "p_value_holm": 0.005997001499250375,
        "point_difference": 0.03620591936690898,
        "status": "reported",
        "valid_bootstrap_iterations": 2000
      },
      {
        "comparison_id": "h1_self_confidence_minus_random_symmetric_seed_404",
        "interval_95": [
          0.28005035470547107,
          0.3051288036465026
        ],
        "method_a": "self_confidence",
        "method_b": "random_review",
        "metric": "average_precision",
        "p_value_holm": 0.005997001499250375,
        "point_difference": 0.2929823051578709,
        "status": "reported",
        "valid_bootstrap_iterations": 2000
      },
      {
        "comparison_id": "h1_self_confidence_minus_random_symmetric_seed_405",
        "interval_95": [
          0.2855317950413325,
          0.31029800350406894
        ],
        "method_a": "self_confidence",
        "method_b": "random_review",
        "metric": "average_precision",
        "p_value_holm": 0.005997001499250375,
        "point_difference": 0.2980656760720609,
        "status": "reported",
        "valid_bootstrap_iterations": 2000
      },
      {
        "comparison_id": "h1_self_confidence_minus_random_symmetric_seed_406",
        "interval_95": [
          0.28864598318220414,
          0.3147144142775111
        ],
        "method_a": "self_confidence",
        "method_b": "random_review",
        "metric": "average_precision",
        "p_value_holm": 0.005997001499250375,
        "point_difference": 0.3017916948127889,
        "status": "reported",
        "valid_bootstrap_iterations": 2000
      },
      {
        "comparison_id": "h3_symmetric_minus_confusion_seed_404",
        "interval_95": [
          9.903678396362704e-05,
          0.016694072333639492
        ],
        "method_a": "self_confidence",
        "method_b": "self_confidence",
        "metric": "average_precision",
        "p_value_holm": 0.047976011994003,
        "point_difference": 0.00859537371850344,
        "status": "reported",
        "valid_bootstrap_iterations": 2000
      },
      {
        "comparison_id": "h3_symmetric_minus_confusion_seed_405",
        "interval_95": [
          -0.0008920853788041305,
          0.014935145257006628
        ],
        "method_a": "self_confidence",
        "method_b": "self_confidence",
        "metric": "average_precision",
        "p_value_holm": 0.047976011994003,
        "point_difference": 0.007070548911689956,
        "status": "reported",
        "valid_bootstrap_iterations": 2000
      },
      {
        "comparison_id": "h3_symmetric_minus_confusion_seed_406",
        "interval_95": [
          0.0028225339332107135,
          0.01958340502470565
        ],
        "method_a": "self_confidence",
        "method_b": "self_confidence",
        "metric": "average_precision",
        "p_value_holm": 0.010494752623688156,
        "point_difference": 0.011084901671939262,
        "status": "reported",
        "valid_bootstrap_iterations": 2000
      },
      {
        "comparison_id": "h3_symmetric_minus_instance_dependent_seed_404",
        "interval_95": [
          0.2425216590631316,
          0.27018943622263336
        ],
        "method_a": "self_confidence",
        "method_b": "self_confidence",
        "metric": "average_precision",
        "p_value_holm": 0.0029985007496251873,
        "point_difference": 0.2567751479070569,
        "status": "reported",
        "valid_bootstrap_iterations": 2000
      },
      {
        "comparison_id": "h3_symmetric_minus_instance_dependent_seed_405",
        "interval_95": [
          0.24821729189853775,
          0.27501767585215786
        ],
        "method_a": "self_confidence",
        "method_b": "self_confidence",
        "metric": "average_precision",
        "p_value_holm": 0.0029985007496251873,
        "point_difference": 0.2617884234318757,
        "status": "reported",
        "valid_bootstrap_iterations": 2000
      },
      {
        "comparison_id": "h3_symmetric_minus_instance_dependent_seed_406",
        "interval_95": [
          0.251197699867191,
          0.2792514720122364
        ],
        "method_a": "self_confidence",
        "method_b": "self_confidence",
        "metric": "average_precision",
        "p_value_holm": 0.0029985007496251873,
        "point_difference": 0.2654366020149367,
        "status": "reported",
        "valid_bootstrap_iterations": 2000
      },
      {
        "comparison_id": "h5_hybrid_minus_self_confidence_confusion_seed_404",
        "interval_95": [
          0.04670013809694134,
          0.05853389470730373
        ],
        "method_a": "fixed_hybrid",
        "method_b": "self_confidence",
        "metric": "average_precision",
        "p_value_holm": 0.005997001499250375,
        "point_difference": 0.052547546590662686,
        "status": "reported",
        "valid_bootstrap_iterations": 2000
      },
      {
        "comparison_id": "h5_hybrid_minus_self_confidence_confusion_seed_405",
        "interval_95": [
          0.04848609419495226,
          0.06106956525861538
        ],
        "method_a": "fixed_hybrid",
        "method_b": "self_confidence",
        "metric": "average_precision",
        "p_value_holm": 0.005997001499250375,
        "point_difference": 0.054609070595176445,
        "status": "reported",
        "valid_bootstrap_iterations": 2000
      },
      {
        "comparison_id": "h5_hybrid_minus_self_confidence_confusion_seed_406",
        "interval_95": [
          0.04810656150942174,
          0.06049983926275598
        ],
        "method_a": "fixed_hybrid",
        "method_b": "self_confidence",
        "metric": "average_precision",
        "p_value_holm": 0.005997001499250375,
        "point_difference": 0.05410428232923786,
        "status": "reported",
        "valid_bootstrap_iterations": 2000
      },
      {
        "comparison_id": "h5_hybrid_minus_self_confidence_group_conditional_seed_404",
        "interval_95": [
          0.05380894560526866,
          0.06675311256205106
        ],
        "method_a": "fixed_hybrid",
        "method_b": "self_confidence",
        "metric": "average_precision",
        "p_value_holm": 0.005997001499250375,
        "point_difference": 0.06026788217486534,
        "status": "reported",
        "valid_bootstrap_iterations": 2000
      },
      {
        "comparison_id": "h5_hybrid_minus_self_confidence_group_conditional_seed_405",
        "interval_95": [
          0.05875182357682752,
          0.07145685380983564
        ],
        "method_a": "fixed_hybrid",
        "method_b": "self_confidence",
        "metric": "average_precision",
        "p_value_holm": 0.005997001499250375,
        "point_difference": 0.06513466434335557,
        "status": "reported",
        "valid_bootstrap_iterations": 2000
      },
      {
        "comparison_id": "h5_hybrid_minus_self_confidence_group_conditional_seed_406",
        "interval_95": [
          0.05899226976700777,
          0.07187112377725746
        ],
        "method_a": "fixed_hybrid",
        "method_b": "self_confidence",
        "metric": "average_precision",
        "p_value_holm": 0.005997001499250375,
        "point_difference": 0.06550191383064385,
        "status": "reported",
        "valid_bootstrap_iterations": 2000
      },
      {
        "comparison_id": "h5_hybrid_minus_self_confidence_instance_dependent_seed_404",
        "interval_95": [
          0.01848300345686238,
          0.023472166038669348
        ],
        "method_a": "fixed_hybrid",
        "method_b": "self_confidence",
        "metric": "average_precision",
        "p_value_holm": 0.005997001499250375,
        "point_difference": 0.020983062831271382,
        "status": "reported",
        "valid_bootstrap_iterations": 2000
      },
      {
        "comparison_id": "h5_hybrid_minus_self_confidence_instance_dependent_seed_405",
        "interval_95": [
          0.01848300345686238,
          0.023472166038669348
        ],
        "method_a": "fixed_hybrid",
        "method_b": "self_confidence",
        "metric": "average_precision",
        "p_value_holm": 0.005997001499250375,
        "point_difference": 0.020983062831271382,
        "status": "reported",
        "valid_bootstrap_iterations": 2000
      },
      {
        "comparison_id": "h5_hybrid_minus_self_confidence_instance_dependent_seed_406",
        "interval_95": [
          0.01848300345686238,
          0.023472166038669348
        ],
        "method_a": "fixed_hybrid",
        "method_b": "self_confidence",
        "metric": "average_precision",
        "p_value_holm": 0.005997001499250375,
        "point_difference": 0.020983062831271382,
        "status": "reported",
        "valid_bootstrap_iterations": 2000
      },
      {
        "comparison_id": "h5_hybrid_minus_self_confidence_symmetric_seed_404",
        "interval_95": [
          0.056457191895439775,
          0.06895618564219294
        ],
        "method_a": "fixed_hybrid",
        "method_b": "self_confidence",
        "metric": "average_precision",
        "p_value_holm": 0.005997001499250375,
        "point_difference": 0.06259236737429391,
        "status": "reported",
        "valid_bootstrap_iterations": 2000
      },
      {
        "comparison_id": "h5_hybrid_minus_self_confidence_symmetric_seed_405",
        "interval_95": [
          0.057757172368690685,
          0.07085607136808317
        ],
        "method_a": "fixed_hybrid",
        "method_b": "self_confidence",
        "metric": "average_precision",
        "p_value_holm": 0.005997001499250375,
        "point_difference": 0.06436561345266478,
        "status": "reported",
        "valid_bootstrap_iterations": 2000
      },
      {
        "comparison_id": "h5_hybrid_minus_self_confidence_symmetric_seed_406",
        "interval_95": [
          0.058360870716253524,
          0.07121816417089324
        ],
        "method_a": "fixed_hybrid",
        "method_b": "self_confidence",
        "metric": "average_precision",
        "p_value_holm": 0.005997001499250375,
        "point_difference": 0.06465623331382025,
        "status": "reported",
        "valid_bootstrap_iterations": 2000
      },
      {
        "comparison_id": "h6_pathology_minus_imagenet_seed_404",
        "interval_95": null,
        "method_a": "self_confidence",
        "method_b": "self_confidence",
        "metric": "average_precision",
        "p_value_holm": null,
        "point_difference": null,
        "status": "not_available_frozen_optional_cell",
        "valid_bootstrap_iterations": 0
      },
      {
        "comparison_id": "h6_pathology_minus_imagenet_seed_405",
        "interval_95": null,
        "method_a": "self_confidence",
        "method_b": "self_confidence",
        "metric": "average_precision",
        "p_value_holm": null,
        "point_difference": null,
        "status": "not_available_frozen_optional_cell",
        "valid_bootstrap_iterations": 0
      },
      {
        "comparison_id": "h6_pathology_minus_imagenet_seed_406",
        "interval_95": null,
        "method_a": "self_confidence",
        "method_b": "self_confidence",
        "metric": "average_precision",
        "p_value_holm": null,
        "point_difference": null,
        "status": "not_available_frozen_optional_cell",
        "valid_bootstrap_iterations": 0
      },
      {
        "comparison_id": "h7_highlighted_minus_context_seed_404",
        "interval_95": [
          -0.005061713152329643,
          0.010786598335651838
        ],
        "method_a": "self_confidence",
        "method_b": "self_confidence",
        "metric": "average_precision",
        "p_value_holm": 0.7226386806596701,
        "point_difference": 0.0029966604528455787,
        "status": "reported",
        "valid_bootstrap_iterations": 2000
      },
      {
        "comparison_id": "h7_highlighted_minus_context_seed_405",
        "interval_95": [
          -0.00817522462518073,
          0.008265313008360732
        ],
        "method_a": "self_confidence",
        "method_b": "self_confidence",
        "metric": "average_precision",
        "p_value_holm": 0.9645177411294353,
        "point_difference": -3.9385652547130334e-06,
        "status": "reported",
        "valid_bootstrap_iterations": 2000
      },
      {
        "comparison_id": "h7_highlighted_minus_context_seed_406",
        "interval_95": [
          -0.013518191519256529,
          0.00313028778773272
        ],
        "method_a": "self_confidence",
        "method_b": "self_confidence",
        "metric": "average_precision",
        "p_value_holm": 0.9645177411294353,
        "point_difference": -0.005371178299694024,
        "status": "reported",
        "valid_bootstrap_iterations": 2000
      }
    ],
    "completed_required_cells": 185,
    "completion_evidence_sha256": "77a02877f9882a608968303e077963abdfbcfabdc997424cfaceadf0df86349c",
    "failed_required_cells": 0,
    "h2_subgroups": {
      "dimensions": {
        "class": {
          "reported_average_precision_range": [
            0.013587049419532534,
            0.8551460145592114
          ],
          "reported_count": 6300,
          "row_count": 6475,
          "status_counts": {
            "not_applicable_zero_corruption": 175,
            "reported": 6300
          }
        },
        "mechanism": {
          "reported_average_precision_range": [
            0.03305238318198219,
            0.642851682089897
          ],
          "reported_count": 1260,
          "row_count": 1295,
          "status_counts": {
            "not_applicable_zero_corruption": 35,
            "reported": 1260
          }
        },
        "rate": {
          "reported_average_precision_range": [
            0.03305238318198219,
            0.642851682089897
          ],
          "reported_count": 1260,
          "row_count": 1295,
          "status_counts": {
            "not_applicable_zero_corruption": 35,
            "reported": 1260
          }
        },
        "tissue": {
          "reported_average_precision_range": [
            0.018489098952010445,
            0.7712351634726737
          ],
          "reported_count": 23940,
          "row_count": 24605,
          "status_counts": {
            "not_applicable_zero_corruption": 665,
            "reported": 23940
          }
        }
      },
      "interpretation_scope": "descriptive_heterogeneity_not_an_omnibus_causal_test",
      "reported_count": 32760,
      "row_count": 33670,
      "source_path": "primary_subgroups.csv",
      "source_sha256": "36be649fef067de82cd11b77f508f0a6fe62f649d393a1c9975a4523c24d166e"
    },
    "h4_restoration": {
      "audit_guided_macro_f1": 0.5244310209334606,
      "cell": {
        "cell_id": "primary_0027_8531672acd3c",
        "classifier_id": "multinomial_logistic_regression",
        "corruption_seed": 404,
        "mechanism": "symmetric_random_corruption",
        "rate": 0.1,
        "representation_id": "imagenet_resnet18_highlighted",
        "required": true,
        "scenario_id": "scenario_0004_97704bd2905b"
      },
      "comparison_id": "audit_guided_minus_random_macro_f1",
      "corrupted_observed_baseline_macro_f1": 0.5267165399263279,
      "directional_result": "adverse_to_registered_hypothesis",
      "interval_95": [
        -0.0028586383107464695,
        -0.0013925186647962915
      ],
      "metric": "macro_f1",
      "point_difference": -0.0021560596665870513,
      "probability_positive": 0.0,
      "random_repetitions": 100,
      "random_review_macro_f1_mean": 0.5265870806000477,
      "registered_hypothesis_supported": false,
      "source_path": "restorations/primary_0027_8531672acd3c/restoration.json",
      "source_sha256": "1dd6f9e105066d0ce0314839783d15d3d0a4f96ece7430216bc3c6733961a27e",
      "status": "reported",
      "uncorrupted_reference_baseline_macro_f1": 0.5374994322248652
    },
    "hypothesis_comparisons": {
      "h1": {
        "comparison_count": 12,
        "holm_one_sided_below_0_05_count": 12,
        "interval_crosses_zero_count": 0,
        "interval_excludes_zero_in_positive_direction_count": 12,
        "point_difference_range": [
          0.03620591936690898,
          0.30188350175581824
        ],
        "positive_point_difference_count": 12,
        "reported_count": 12,
        "status_counts": {
          "reported": 12
        }
      },
      "h3": {
        "comparison_count": 6,
        "holm_one_sided_below_0_05_count": 6,
        "interval_crosses_zero_count": 1,
        "interval_excludes_zero_in_positive_direction_count": 5,
        "point_difference_range": [
          0.007070548911689956,
          0.2654366020149367
        ],
        "positive_point_difference_count": 6,
        "reported_count": 6,
        "status_counts": {
          "reported": 6
        }
      },
      "h5": {
        "comparison_count": 12,
        "holm_one_sided_below_0_05_count": 12,
        "interval_crosses_zero_count": 0,
        "interval_excludes_zero_in_positive_direction_count": 12,
        "point_difference_range": [
          0.020983062831271382,
          0.06550191383064385
        ],
        "positive_point_difference_count": 12,
        "reported_count": 12,
        "status_counts": {
          "reported": 12
        }
      },
      "h6": {
        "comparison_count": 3,
        "holm_one_sided_below_0_05_count": 0,
        "interval_crosses_zero_count": 0,
        "interval_excludes_zero_in_positive_direction_count": 0,
        "point_difference_range": null,
        "positive_point_difference_count": 0,
        "reported_count": 0,
        "status_counts": {
          "not_available_frozen_optional_cell": 3
        }
      },
      "h7": {
        "comparison_count": 3,
        "holm_one_sided_below_0_05_count": 0,
        "interval_crosses_zero_count": 3,
        "interval_excludes_zero_in_positive_direction_count": 0,
        "point_difference_range": [
          -0.005371178299694024,
          0.0029966604528455787
        ],
        "positive_point_difference_count": 1,
        "reported_count": 3,
        "status_counts": {
          "reported": 3
        }
      }
    },
    "inference": {
      "interval_label": "95_percentile_bootstrap_interval",
      "multiple_comparison_method": "holm",
      "p_value_definition": "(1 + count(bootstrap_difference <= 0)) / (1 + valid_iterations)",
      "p_value_sidedness": "one_sided"
    },
    "instance_dependent_seed_audit": {
      "disclosure_required": true,
      "independent_corruption_realisations": false,
      "oof_outputs_byte_identical": true,
      "ranking_outputs_byte_identical": true,
      "records": [
        {
          "cell_id": "primary_0187_53201213be77",
          "oof_predictions_sha256": "730459794225e338834568d4d7bf574b65c38b634726e787d9121f818d702958",
          "ranking_sha256": "69766d68a24679b510b3e63e0eb039207119ff1d96953dafe31cfa51f74d276b",
          "seed": 404
        },
        {
          "cell_id": "primary_0193_afae7a244238",
          "oof_predictions_sha256": "730459794225e338834568d4d7bf574b65c38b634726e787d9121f818d702958",
          "ranking_sha256": "69766d68a24679b510b3e63e0eb039207119ff1d96953dafe31cfa51f74d276b",
          "seed": 405
        },
        {
          "cell_id": "primary_0199_6029f5261b38",
          "oof_predictions_sha256": "730459794225e338834568d4d7bf574b65c38b634726e787d9121f818d702958",
          "ranking_sha256": "69766d68a24679b510b3e63e0eb039207119ff1d96953dafe31cfa51f74d276b",
          "seed": 406
        }
      ],
      "reporting_policy": "retain_frozen_rows_but_treat_as_one_deterministic_realisation",
      "seeds": [
        404,
        405,
        406
      ]
    },
    "outcomes_inspected": true,
    "run_id": "20260727T133947.089370Z_pannuke_primary_orphan_recovery",
    "skipped_optional_cells": 37,
    "stage_attestation_record_sha256": "5af827544502fbdf688a73916ec58b5dac0984c5a682a33ce6dfc97538228871",
    "stage_attestation_verification_sha256": "e19f17b14a375e8e048e1d2fc3060c19831e4259c8311831d88f75753024c768",
    "statistics_sha256": "c3685fe9863fd73b1298f0558212cb5267b07c3ce6e4e4f37018dec55c115ac0"
  },
  "primary_study_status": "PRIMARY_STUDY_COMPLETE",
  "publication_limits": {
    "puma_downstream_intervention": "flag_exclude_top_5_percent_training_rows",
    "puma_first_public_combined_commit": "c5bd44193b2abd67bc7e7f1bd9384aa87435d500",
    "puma_flagged_rows_expert_reviewed_or_relabelled": false,
    "puma_official_hidden_challenge_test_used": false,
    "puma_partition": "aanca_defined_144_development_62_final_of_206_public_rois",
    "puma_public_preoutcome_timestamp_available": false,
    "puma_verifier_is_third_party_validation": false,
    "puma_verifier_retrains_models_from_images": false,
    "puma_verifier_scope": "saved_evidence_readback_with_maintained_helpers"
  },
  "realism_stress": {
    "all_class_safeguards_passed_count": 1,
    "all_scenarios_passed": false,
    "candidate_changed": false,
    "claim_boundary": {
      "automatic_annotation_change_permitted": false,
      "clinical_utility_proven": false,
      "independent_confirmation": false,
      "natural_error_detection_proven": false,
      "pathologist_error_detection_proven": false
    },
    "disposition": "post_confirmation_exploratory_stress_only",
    "positive_aggregate_lower_bound_count": 9,
    "scenario_count": 9,
    "source": {
      "path": "artifacts/puma_realism_stress/results.json",
      "sha256": "0091c14075e7304e0a6effef5a398c32c692f40aaefe7ee7d0b126a101cf7892",
      "size_bytes": 45553
    },
    "study_id": "puma_realism_stress_v1"
  },
  "schema_version": 3,
  "scientific_status": "EXTERNAL_VALIDATION_COMPLETE",
  "source_files": {
    "findings_story_script_asset": {
      "path": "assets/findings-evidence-story.js",
      "sha256": "e13625e9d4f34bc90f2ea97c3201dd246a5f40b97f475e6bc8b19c02e61878b8",
      "size_bytes": 3647
    },
    "hero_nucleus_assets": [
      {
        "path": "assets/hero/nuclei/nucleus-compact.png",
        "sha256": "fa5f3f61bba6d509fbb923efff7c160a04204a7d4b00d8ac712f8d505d9eec45",
        "size_bytes": 1570355
      },
      {
        "path": "assets/hero/nuclei/nucleus-elongated.png",
        "sha256": "17f5d449d82d604e01b91f19ae3ac876b1584523dc729e674527a289346d3866",
        "size_bytes": 1708042
      },
      {
        "path": "assets/hero/nuclei/nucleus-kidney.png",
        "sha256": "3ed26fe1475011b8acc5b3447039eb7642f609633b08f94d8706e0ae84494482",
        "size_bytes": 1444775
      },
      {
        "path": "assets/hero/nuclei/nucleus-bilobed.png",
        "sha256": "6f3547e9ca873d3b436c30829983f3f8b75cdef71e2d13feb70b8cd1ec5fac33",
        "size_bytes": 1222501
      },
      {
        "path": "assets/hero/nuclei/nucleus-irregular.png",
        "sha256": "142f75900fccd1242cc063b1c0ee4d02ed97ddaf3efa65e21ef629fdaa397b83",
        "size_bytes": 1284733
      },
      {
        "path": "assets/hero/nuclei/nucleus-flattened.png",
        "sha256": "354b5f9accbddc2b7188abe65ba97cd008d52e71dab05bec1fb3b8174268ab5b",
        "size_bytes": 1805225
      }
    ],
    "method_workflow_asset": {
      "path": "assets/method-audit-workflow.png",
      "sha256": "aaf64096acee49cdcdc0491c4a5aa3749c19912f18b96e3e4b7d88db2c43ad37",
      "size_bytes": 826713
    },
    "primary_run": "artifacts/runs/20260727T133947.089370Z_pannuke_primary_orphan_recovery",
    "qc_bundle": "reports/pannuke_qc",
    "qc_bundle_manifest_sha256": "0b188ecc586ed772b29845e15e169fb492ed8d2ad0f5b1a6643531ccee10857f",
    "qc_overlay_sha256": "a1bd87dd397417d711d1d4937429eae5f5d972d3fa6ffa27a45129339587f10a",
    "released_evidence": {
      "monusac_external": {
        "path": "artifacts/monusac_external_validation/results.json",
        "sha256": "b2724e3e0baedcd0f1eb0fc7dfae127bf3789b03ac626d2701477ff4bae8e7d4",
        "size_bytes": 33249
      },
      "nucls_external": {
        "path": "artifacts/nucls_external_validation/unbiased-v1/results.json",
        "sha256": "a931100a57e8d4b2a34a0216f047de8aa9d21c7275bc1beafd3b361fd955e9f6",
        "size_bytes": 19653
      },
      "nucls_qc_feasibility": {
        "path": "artifacts/nucls_supervised_qc_feasibility/results.json",
        "sha256": "e95dd1416ad2be08126aa710a4d00ff999a4cb25936fcf1f9fc1f71f62d36229",
        "size_bytes": 2949
      },
      "puma_confirmation": {
        "path": "artifacts/puma_new_data_confirmation/results.json",
        "sha256": "e547d77a2ae5bffdaed9f64f83c369292fe11cd2e8a305d02fc3d2271131dd92",
        "size_bytes": 42958
      },
      "puma_label_sensitivity": {
        "path": "artifacts/puma_audit_time_label_sensitivity/results.json",
        "sha256": "8f524b236995a495048a0955ebf930e14e732a1214d3856d3711571af13fd5cd",
        "size_bytes": 9944
      },
      "puma_realism_stress": {
        "path": "artifacts/puma_realism_stress/results.json",
        "sha256": "0091c14075e7304e0a6effef5a398c32c692f40aaefe7ee7d0b126a101cf7892",
        "size_bytes": 45553
      },
      "puma_verification": {
        "path": "artifacts/puma_new_data_confirmation/verification.json",
        "sha256": "2f4b6d6c64d8dfa493a89be0111ca8752257dabce808d4d9fa35933cdd9c5d68",
        "size_bytes": 1268
      }
    }
  }
}