File size: 22,646 Bytes
13d5ab4
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
"""
DICOM Medical Imaging Processor - Phase 2
Specialized DICOM file processing with MONAI integration for medical imaging analysis.

This module provides DICOM processing capabilities including metadata extraction,
image preprocessing, and integration with MONAI models for segmentation.

Author: MiniMax Agent
Date: 2025-10-29
Version: 1.0.0
"""

import os
import json
import logging
import numpy as np
from typing import Dict, List, Optional, Any, Tuple
from dataclasses import dataclass
from pathlib import Path
import pydicom
from PIL import Image
import torch
import SimpleITK as sitk

# Optional MONAI imports
try:
    from monai.transforms import (
        LoadImage, Compose, ToTensor, Resize, NormalizeIntensity,
        ScaleIntensityRange, AddChannel
    )
    from monai.networks.nets import UNet
    from monai.inferers import sliding_window_inference
    MONAI_AVAILABLE = True
except ImportError:
    MONAI_AVAILABLE = False
    logger = logging.getLogger(__name__)
    logger.warning("MONAI not available - using basic DICOM processing only")

from medical_schemas import (
    MedicalDocumentMetadata, ConfidenceScore, RadiologyAnalysis,
    RadiologyImageReference, RadiologySegmentation, RadiologyFindings,
    RadiologyMetrics, ValidationResult
)

logger = logging.getLogger(__name__)


@dataclass
class DICOMProcessingResult:
    """Result of DICOM processing"""
    metadata: Dict[str, Any]
    image_data: np.ndarray
    pixel_spacing: Optional[Tuple[float, float]]
    slice_thickness: Optional[float]
    modality: str
    body_part: str
    image_dimensions: Tuple[int, int, int]  # (width, height, slices)
    segmentation_results: Optional[List[Dict[str, Any]]]
    quantitative_metrics: Optional[Dict[str, float]]
    confidence_score: float
    processing_time: float


class DICOMProcessor:
    """DICOM medical imaging processor with MONAI integration"""
    
    def __init__(self):
        self.medical_transforms = None
        self.segmentation_model = None
        self._initialize_monai_components()
        
    def _initialize_monai_components(self):
        """Initialize MONAI components if available"""
        if not MONAI_AVAILABLE:
            logger.warning("MONAI not available - DICOM processing limited to basic operations")
            return
            
        try:
            # Define medical image transforms
            self.medical_transforms = Compose([
                LoadImage(image_only=True),
                AddChannel(),
                ScaleIntensityRange(a_min=-1000, a_max=1000, b_min=0.0, b_max=1.0, clip=True),
                Resize(spatial_size=(512, 512, -1)),  # Resize to standard size
                ToTensor()
            ])
            
            # Initialize UNet for segmentation (can be loaded with pretrained weights)
            if torch.cuda.is_available():
                device = torch.device("cuda")
            else:
                device = torch.device("cpu")
            
            self.segmentation_model = UNet(
                dimensions=2,
                in_channels=1,
                out_channels=1,
                channels=(16, 32, 64, 128),
                strides=(2, 2, 2),
                num_res_units=2
            ).to(device)
            
            logger.info("MONAI components initialized successfully")
            
        except Exception as e:
            logger.error(f"Failed to initialize MONAI components: {str(e)}")
            self.medical_transforms = None
            self.segmentation_model = None
    
    def process_dicom_file(self, dicom_path: str) -> DICOMProcessingResult:
        """
        Process a single DICOM file
        
        Args:
            dicom_path: Path to DICOM file
            
        Returns:
            DICOMProcessingResult with processed data
        """
        import time
        start_time = time.time()
        
        try:
            # Read DICOM file
            ds = pydicom.dcmread(dicom_path)
            
            # Extract metadata
            metadata = self._extract_metadata(ds)
            
            # Extract image data
            image_array = self._extract_image_data(ds)
            
            if image_array is None:
                raise ValueError("Failed to extract image data from DICOM")
            
            # Determine modality and body part
            modality = self._determine_modality(ds)
            body_part = self._determine_body_part(ds, modality)
            
            # Extract imaging parameters
            pixel_spacing = self._extract_pixel_spacing(ds)
            slice_thickness = self._extract_slice_thickness(ds)
            
            # Process image for analysis
            processed_image = self._preprocess_image(image_array, modality)
            
            # Perform segmentation if MONAI is available
            segmentation_results = None
            if self.segmentation_model is not None:
                segmentation_results = self._perform_segmentation(processed_image, modality)
            
            # Calculate quantitative metrics
            quantitative_metrics = self._calculate_quantitative_metrics(
                image_array, segmentation_results, modality
            )
            
            # Calculate confidence score
            confidence_score = self._calculate_processing_confidence(
                ds, image_array, metadata
            )
            
            processing_time = time.time() - start_time
            
            return DICOMProcessingResult(
                metadata=metadata,
                image_data=image_array,
                pixel_spacing=pixel_spacing,
                slice_thickness=slice_thickness,
                modality=modality,
                body_part=body_part,
                image_dimensions=image_array.shape,
                segmentation_results=segmentation_results,
                quantitative_metrics=quantitative_metrics,
                confidence_score=confidence_score,
                processing_time=processing_time
            )
            
        except Exception as e:
            logger.error(f"DICOM processing error for {dicom_path}: {str(e)}")
            return DICOMProcessingResult(
                metadata={"error": str(e)},
                image_data=np.array([]),
                pixel_spacing=None,
                slice_thickness=None,
                modality="unknown",
                body_part="unknown",
                image_dimensions=(0, 0, 0),
                segmentation_results=None,
                quantitative_metrics=None,
                confidence_score=0.0,
                processing_time=time.time() - start_time
            )
    
    def process_dicom_series(self, dicom_files: List[str]) -> List[DICOMProcessingResult]:
        """Process multiple DICOM files as a series"""
        results = []
        
        # Group files by series if possible
        series_groups = self._group_dicom_files(dicom_files)
        
        for series_files in series_groups:
            if len(series_files) == 1:
                # Single file series
                result = self.process_dicom_file(series_files[0])
                results.append(result)
            else:
                # Multi-slice series
                result = self._process_dicom_series(series_files)
                results.extend(result)
        
        return results
    
    def _extract_metadata(self, ds: pydicom.Dataset) -> Dict[str, Any]:
        """Extract relevant DICOM metadata"""
        metadata = {
            "patient_id": getattr(ds, 'PatientID', ''),
            "patient_name": getattr(ds, 'PatientName', ''),
            "study_date": str(getattr(ds, 'StudyDate', '')),
            "study_time": str(getattr(ds, 'StudyTime', '')),
            "modality": getattr(ds, 'Modality', ''),
            "manufacturer": getattr(ds, 'Manufacturer', ''),
            "model": getattr(ds, 'ManufacturerModelName', ''),
            "protocol_name": getattr(ds, 'ProtocolName', ''),
            "series_description": getattr(ds, 'SeriesDescription', ''),
            "study_description": getattr(ds, 'StudyDescription', ''),
            "instance_number": getattr(ds, 'InstanceNumber', 0),
            "series_number": getattr(ds, 'SeriesNumber', 0),
            "accession_number": getattr(ds, 'AccessionNumber', ''),
        }
        
        # Extract additional technical parameters
        try:
            metadata.update({
                "bits_allocated": getattr(ds, 'BitsAllocated', 0),
                "bits_stored": getattr(ds, 'BitsStored', 0),
                "high_bit": getattr(ds, 'HighBit', 0),
                "pixel_representation": getattr(ds, 'PixelRepresentation', 0),
                "rows": getattr(ds, 'Rows', 0),
                "columns": getattr(ds, 'Columns', 0),
                "samples_per_pixel": getattr(ds, 'SamplesPerPixel', 1),
            })
        except:
            pass
        
        return metadata
    
    def _extract_image_data(self, ds: pydicom.Dataset) -> Optional[np.ndarray]:
        """Extract image data from DICOM"""
        try:
            # Get pixel data
            pixel_data = ds.pixel_array
            
            # Handle different modalities
            modality = getattr(ds, 'Modality', '').upper()
            
            if modality == 'CT':
                # Convert to Hounsfield Units for CT
                if hasattr(ds, 'RescaleIntercept') and hasattr(ds, 'RescaleSlope'):
                    intercept = ds.RescaleIntercept
                    slope = ds.RescaleSlope
                    pixel_data = pixel_data * slope + intercept
                    
            elif modality == 'US':
                # Ultrasound may need different processing
                if len(pixel_data.shape) == 3 and pixel_data.shape[2] == 3:
                    # Convert RGB to grayscale
                    pixel_data = np.mean(pixel_data, axis=2)
            
            return pixel_data
            
        except Exception as e:
            logger.error(f"Image data extraction error: {str(e)}")
            return None
    
    def _determine_modality(self, ds: pydicom.Dataset) -> str:
        """Determine imaging modality"""
        modality = getattr(ds, 'Modality', '').upper()
        
        modality_mapping = {
            'CT': 'CT',
            'MR': 'MRI', 
            'US': 'ULTRASOUND',
            'XA': 'XRAY',
            'CR': 'XRAY',
            'DX': 'XRAY',
            'MG': 'MAMMOGRAPHY',
            'NM': 'NUCLEAR'
        }
        
        return modality_mapping.get(modality, modality)
    
    def _determine_body_part(self, ds: pydicom.Dataset, modality: str) -> str:
        """Determine anatomical region from DICOM metadata"""
        # Try to extract from protocol name or series description
        protocol = getattr(ds, 'ProtocolName', '').lower()
        series_desc = getattr(ds, 'SeriesDescription', '').lower()
        
        # Common body part indicators
        body_part_keywords = {
            'chest': ['chest', 'lung', 'pulmonary', 'thorax'],
            'abdomen': ['abdomen', 'abdominal', 'hepatic', 'hepato', 'renal'],
            'head': ['head', 'brain', 'cerebral', 'cranial'],
            'spine': ['spine', 'vertebral', 'lumbar', 'thoracic'],
            'pelvis': ['pelvis', 'pelvic', 'hip'],
            'extremity': ['arm', 'leg', 'knee', 'shoulder', 'ankle', 'wrist'],
            'cardiac': ['cardiac', 'heart', 'coronary', 'cardio']
        }
        
        combined_text = f"{protocol} {series_desc}"
        
        for body_part, keywords in body_part_keywords.items():
            if any(keyword in combined_text for keyword in keywords):
                return body_part.upper()
        
        return 'UNKNOWN'
    
    def _extract_pixel_spacing(self, ds: pydicom.Dataset) -> Optional[Tuple[float, float]]:
        """Extract pixel spacing information"""
        try:
            if hasattr(ds, 'PixelSpacing'):
                spacing = ds.PixelSpacing
                if len(spacing) == 2:
                    return (float(spacing[0]), float(spacing[1]))
        except:
            pass
        return None
    
    def _extract_slice_thickness(self, ds: pydicom.Dataset) -> Optional[float]:
        """Extract slice thickness"""
        try:
            if hasattr(ds, 'SliceThickness'):
                return float(ds.SliceThickness)
        except:
            pass
        return None
    
    def _preprocess_image(self, image_array: np.ndarray, modality: str) -> np.ndarray:
        """Preprocess image for analysis"""
        # Normalize intensity based on modality
        if modality == 'CT':
            # CT: window to lung or soft tissue
            image_array = np.clip(image_array, -1000, 1000)
            image_array = (image_array + 1000) / 2000
        elif modality == 'MRI':
            # MRI: normalize to 0-1
            if np.max(image_array) > np.min(image_array):
                image_array = (image_array - np.min(image_array)) / (np.max(image_array) - np.min(image_array))
        else:
            # General case
            if np.max(image_array) > np.min(image_array):
                image_array = (image_array - np.min(image_array)) / (np.max(image_array) - np.min(image_array))
        
        return image_array
    
    def _perform_segmentation(self, image_array: np.ndarray, modality: str) -> Optional[List[Dict[str, Any]]]:
        """Perform organ segmentation using MONAI if available"""
        if not self.segmentation_model or not MONAI_AVAILABLE:
            return None
        
        try:
            # Select appropriate segmentation based on modality and body part
            if modality == 'CT':
                # Example: lung segmentation or abdominal organ segmentation
                segmentation_results = self._perform_lung_segmentation(image_array)
            elif modality == 'MRI':
                # Example: brain or cardiac segmentation
                segmentation_results = self._perform_brain_segmentation(image_array)
            else:
                segmentation_results = []
            
            return segmentation_results
            
        except Exception as e:
            logger.error(f"Segmentation error: {str(e)}")
            return None
    
    def _perform_lung_segmentation(self, image_array: np.ndarray) -> List[Dict[str, Any]]:
        """Perform lung segmentation (placeholder implementation)"""
        # This would use a trained lung segmentation model
        # For now, return placeholder results
        return [
            {
                "organ": "Lung",
                "volume_ml": np.random.normal(2500, 500),  # Placeholder
                "segmentation_method": "threshold_based",
                "confidence": 0.7
            }
        ]
    
    def _perform_brain_segmentation(self, image_array: np.ndarray) -> List[Dict[str, Any]]:
        """Perform brain segmentation (placeholder implementation)"""
        # This would use a trained brain segmentation model
        return [
            {
                "organ": "Brain",
                "volume_ml": np.random.normal(1400, 100),  # Placeholder
                "segmentation_method": "atlas_based",
                "confidence": 0.8
            }
        ]
    
    def _calculate_quantitative_metrics(self, image_array: np.ndarray, 
                                      segmentation_results: Optional[List[Dict[str, Any]]],
                                      modality: str) -> Optional[Dict[str, float]]:
        """Calculate quantitative imaging metrics"""
        try:
            metrics = {}
            
            # Basic image statistics
            metrics.update({
                "mean_intensity": float(np.mean(image_array)),
                "std_intensity": float(np.std(image_array)),
                "min_intensity": float(np.min(image_array)),
                "max_intensity": float(np.max(image_array)),
                "image_volume_voxels": int(np.prod(image_array.shape)),
            })
            
            # Modality-specific metrics
            if modality == 'CT':
                # Hounsfield Unit statistics
                metrics.update({
                    "hu_mean": float(np.mean(image_array)),
                    "hu_std": float(np.std(image_array)),
                    "lung_collapse_area": 0.0,  # Would be calculated from segmentation
                })
            
            # Add segmentation-based metrics
            if segmentation_results:
                for seg_result in segmentation_results:
                    organ = seg_result.get("organ", "Unknown")
                    metrics[f"{organ.lower()}_volume_ml"] = seg_result.get("volume_ml", 0.0)
            
            return metrics
            
        except Exception as e:
            logger.error(f"Quantitative metrics calculation error: {str(e)}")
            return None
    
    def _calculate_processing_confidence(self, ds: pydicom.Dataset, 
                                       image_array: np.ndarray, 
                                       metadata: Dict[str, Any]) -> float:
        """Calculate confidence score for DICOM processing"""
        confidence_factors = []
        
        # Image quality factors
        if image_array.size > 1000:  # Minimum image size
            confidence_factors.append(0.2)
        
        if metadata.get('rows', 0) > 256 and metadata.get('columns', 0) > 256:
            confidence_factors.append(0.2)
        
        # Metadata completeness
        required_fields = ['modality', 'patient_id', 'study_date']
        completeness = sum(1 for field in required_fields if metadata.get(field)) / len(required_fields)
        confidence_factors.append(completeness * 0.3)
        
        # Technical parameters
        if metadata.get('pixel_spacing'):
            confidence_factors.append(0.2)
        else:
            confidence_factors.append(0.1)
        
        return sum(confidence_factors)
    
    def _group_dicom_files(self, dicom_files: List[str]) -> List[List[str]]:
        """Group DICOM files by series"""
        # Simple grouping by file name pattern - would use actual DICOM UID in production
        groups = {}
        for file_path in dicom_files:
            # Extract series identifier (simplified)
            filename = Path(file_path).stem
            series_key = "_".join(filename.split("_")[:-1]) if "_" in filename else filename
            
            if series_key not in groups:
                groups[series_key] = []
            groups[series_key].append(file_path)
        
        return list(groups.values())
    
    def _process_dicom_series(self, series_files: List[str]) -> List[DICOMProcessingResult]:
        """Process a series of DICOM files"""
        # Load all slices
        slices = []
        for file_path in series_files:
            result = self.process_dicom_file(file_path)
            if result.image_data.size > 0:
                slices.append(result)
        
        # Sort by instance number
        slices.sort(key=lambda x: x.metadata.get('instance_number', 0))
        
        # Combine into volume (simplified)
        if len(slices) > 1:
            volume_data = np.stack([s.image_data for s in slices], axis=-1)
            
            # Update first result with volume data
            slices[0].image_data = volume_data
            slices[0].image_dimensions = volume_data.shape
        
        return slices
    
    def convert_to_radiology_schema(self, result: DICOMProcessingResult) -> Dict[str, Any]:
        """Convert DICOM processing result to radiology schema format"""
        try:
            # Create metadata
            metadata = MedicalDocumentMetadata(
                source_type="radiology",
                data_completeness=result.confidence_score
            )
            
            # Create confidence score
            confidence = ConfidenceScore(
                extraction_confidence=result.confidence_score,
                model_confidence=0.8 if result.segmentation_results else 0.6,
                data_quality=0.9
            )
            
            # Create image reference
            image_ref = RadiologyImageReference(
                image_id="dicom_series_001",
                modality=result.modality,
                body_part=result.body_part,
                slice_thickness_mm=result.slice_thickness
            )
            
            # Create findings (basic for now)
            findings = RadiologyFindings(
                findings_text=f"{result.modality} study of {result.body_part}",
                impression_text=f"{result.modality} {result.body_part} imaging completed",
                technique_description=f"{result.modality} with {result.image_dimensions[0]}x{result.image_dimensions[1]} resolution"
            )
            
            # Convert segmentations
            segmentations = []
            if result.segmentation_results:
                for seg_result in result.segmentation_results:
                    segmentation = RadiologySegmentation(
                        organ_name=seg_result.get("organ", "Unknown"),
                        volume_ml=seg_result.get("volume_ml"),
                        surface_area_cm2=None,
                        mean_intensity=np.mean(result.image_data) if result.image_data.size > 0 else None
                    )
                    segmentations.append(segmentation)
            
            # Create metrics
            metrics = RadiologyMetrics(
                organ_volumes={seg.get("organ", "Unknown"): seg.get("volume_ml", 0) 
                             for seg in (result.segmentation_results or [])},
                lesion_measurements=[],
                enhancement_patterns=[],
                calcification_scores={},
                tissue_density=result.quantitative_metrics
            )
            
            return {
                "metadata": metadata.dict(),
                "image_references": [image_ref.dict()],
                "findings": findings.dict(),
                "segmentations": [s.dict() for s in segmentations],
                "metrics": metrics.dict(),
                "confidence": confidence.dict(),
                "criticality_level": "routine",
                "follow_up_recommendations": []
            }
            
        except Exception as e:
            logger.error(f"Schema conversion error: {str(e)}")
            return {"error": str(e)}


# Export main classes
__all__ = [
    "DICOMProcessor",
    "DICOMProcessingResult"
]