Pydicom - Medical Imaging Standards

When to Use

• Processing medical imaging data (CT, MRI, X-ray, ultrasound).
• Extracting patient metadata and clinical information from DICOM files.
• Building AI models for radiology that require both image and metadata.
• Converting DICOM to other formats for analysis.
• Quality assurance and compliance checking in medical imaging workflows.

Core Principles

Datasets as Dicts

ds.PatientName

ds[0x0010, 0x0010]

Pixel Data

PixelData

pixel_array

VR (Value Representation)

Quick Reference

Standard Imports

import pydicom
from pydicom.data import get_testdata_files
import matplotlib.pyplot as plt
import numpy as np

Basic Patterns

# 1. Read file
ds = pydicom.dcmread("scan.dcm")

# 2. Access Metadata
print(f"Patient: {ds.PatientName}, ID: {ds.PatientID}")
print(f"Modality: {ds.Modality}") # CT, MR, DX
print(f"Study Date: {ds.StudyDate}")
print(f"Slice Thickness: {ds.SliceThickness}")

# 3. Access Image
plt.imshow(ds.pixel_array, cmap="gray")
plt.title(f"{ds.Modality} - {ds.PatientName}")

Critical Rules

✅ DO

• Use pixel_array property - Always access pixel data via

  ds.pixel_array

  rather than

  ds.PixelData

  for proper NumPy integration.
• Check for missing tags - Use

  hasattr(ds, 'TagName')

  before accessing optional tags.
• Respect patient privacy - DICOM files contain PHI (Protected Health Information). Always anonymize before sharing.
• Handle different photometric interpretations - Some images may be inverted or use different color spaces.

❌ DON'T

• Don't modify DICOM files in place - Always create a copy when modifying to preserve original data.
• Don't ignore VR types - DICOM has strict data types. Converting incorrectly can corrupt data.
• Don't assume all DICOM files have images - Some contain only metadata (structured reports).

Advanced Patterns

Working with DICOM Series

import pydicom
from pathlib import Path

# Load a series of DICOM files
dicom_dir = Path("dicom_series")
files = sorted(dicom_dir.glob("*.dcm"))

# Load and stack slices
slices = [pydicom.dcmread(f) for f in files]
volume = np.stack([s.pixel_array for s in slices])

Anonymization

# Remove patient identifiers
ds.PatientName = "ANONYMOUS"
ds.PatientID = "000000"
ds.PatientBirthDate = ""
ds.PatientSex = ""