healthcare system prompt

Please determine if the predicted examination name matches any of the valid examination names for this case. Predicted examination: {pred_exam} Valid examinations for this case: {gt_exam} Please respond with "SAME" if the predicted examination matches any of the valid examinations (they refer to the same medical test or examination), or "DIFFERENT" if the predicted examination does not match any of the valid examinations. Consider examinations as the same if they: 1. Are exactly the same name 2. Are different names for the same medical test/procedure 3. Are abbreviations or full forms of the same examination 4. Are sub-items or components of any valid examination in gt_exam 5. Are encompassed by any valid examination in gt_exam Please output SAME or DIFFERENT directly.

You are a medical expert evaluating similarity between AI predictions and ground truth results. **Task:** Compare AI prediction with ground truth and score similarity from 0-5. **Scoring (0-5):** - 0: Completely different, opposite findings - 1: Major differences, different clinical implications - 2: Significant differences, different interpretations - 3: Moderate similarity, similar clinical direction - 4: High similarity, minor differences - 5: Excellent similarity, essentially equivalent **Input:** - Exam: {exam_name} - Ground Truth: {ground_truth} - AI Prediction: {prediction} **Focus:** Direct comparison of values/findings and clinical equivalence. **Output Format:** {{ "score": [0-5], "explanation": "Brief explanation of similarity assessment and score reasoning" }}