Abstract

We predict the BIRADS by applying a Convolutional neural network.

Intelligent BIRADS lights in support of the diagnosis of breast cancer in Baja California.

Introduction

Problem

Since breast cancer is the first dead cause in Mexico among women it became a big public health problem —in fact, 2.26 million cases worldwide. In other words, is a type of cancer with the highest incidence and mortality in women: every day at least 14 women, chiefly between 50 to 69 years, die. Indeed, as we can see in the below figure, breast cancer is an increasing tendency compared to other cancers.

Nowadays, doctors carried out analyses using Traditional 2D mammograms from patient requests at public Mexican hospitals. In Ensenada, doctors request private external assistance. Oncologists annotate medical images, and they chiefly say what is the patient's BI-RADS score. "BI-RADS" means Breast Imaging Reporting and Database System, and it's scoring standard radiologists and oncologists use to describe mammogram results. We'll explain further BI-RADS in the section.

Our goals are to build data mining models that understand mammograms and predict breast cancer developing risk, continuing works. Since our model output is a person's future healthy situation, we'll do descriptive and predictive methods, indeed we're going to apply Machine Learning algorithms to datasets. Of course, we don't expect to replace medical doctors but assist them. We know other computer-aided detection systems have been developed for breast cancer detection but no one applies them to regional cities and they are not free.

We expect our project can help thousands of women in quick cancer detection because data mining is faster and cheaper than humans, therefore we're contributing to the decrease in the death rate.

Related work

@startuml

actor Paciente
actor Recepción
actor TecnólogoRadiología
actor Radiólogo
actor HIS
note left HIS: Hospital Information System
database DICOM
actor Modelo
actor PACS
note left PACS: Picture Archiving and Communication System
actor Oncólogo
actor Proveedor
actor ProgramaDeCancer

Paciente -> Recepción : Solicita mamografía
Recepción -> HIS : Registra solicitud
HIS --> TecnólogoRadiología : Confirma solicitud
TecnólogoRadiología -> Paciente : Realiza mamografía
TecnólogoRadiología -> PACS : Envía imágenes de mamografía
PACS --> TecnólogoRadiología : Confirma recepción de imágenes
PACS --> Proveedor: Emite imágenes a proveedor
PACS -> Radiólogo : Notifica disponibilidad de nuevas imágenes
Proveedor -> PACS : Revisa e interpreta imágenes
PACS -> ProgramaDeCancer: Muestra resultados
ProgramaDeCancer -> Paciente: Emite veridicto sobre biopsia

@enduml

Propuesta

@startuml

actor Paciente
actor Recepción
actor TecnólogoRadiología
actor Radiólogo
actor HIS
note left HIS: Hospital Information System
database DICOM
actor Modelo
actor PACS
note left PACS: Picture Archiving and Communication System
actor Oncólogo
actor Proveedor
actor ProgramaDeCancer

Paciente -> Recepción : Solicita mamografía
Recepción -> HIS : Registra solicitud
HIS --> TecnólogoRadiología : Confirma solicitud
TecnólogoRadiología -> Paciente : Realiza mamografía
TecnólogoRadiología -> PACS : Envía imágenes de mamografía
PACS --> TecnólogoRadiología : Confirma recepción de imágenes
PACS -> Modelo: Emite información del paciente (Mastografías o factores de riesgo)
Modelo -> PACS: Guarda resultado (BIRADS y confianza)
PACS -> Radiólogo: Muestra resultado de Modelo
PACS --> Proveedor: Emite imágenes a proveedor (ordenando por prioridad)
PACS -> Radiólogo : Notifica disponibilidad de nuevas imágenes
Proveedor -> PACS : Revisa e interpreta imágenes
PACS -> ProgramaDeCancer: Muestra resultados
ProgramaDeCancer -> Paciente: Emite veridicto sobre biopsia

@enduml

Methodology

1. Key concepts

2. DICOM VIEW

3. Preprocessing DDSM

4. Engineering process

DICOM

Dicom Web

Development

Functional requirements

Those functional requirements ordered by priority are authenticated users, showing a list of patients, adding patients, adding new Screening Studies, viewing screening tests, confirming cancer diagnosis or not in the screening test, listing screening tests, managing users, and calculating the suspicion of breast cancer.

Imagenes (4 mastografias) en formato DCM almacenadas en carpetas de Windows 1…N cuyo directorio son expendientes ID (no vinculado con el servicio web).

Vincular las carpetas y el expendiente electrónico HTML y CSS y también en formato PDF.

Precesamiento de imagenes con OpenCV.

BIRDS y posibles caracteristicas.

"Expediente ID con mastografía".

graph TD
  Radiologo --> Servidor

Expect results.

CRUD.

Image.

After sending neuronal networks that send a BIRDS and identify the parts, preprocessing by mammograms.

Get BIRDS.

Identify warning zones.

Get interpretation.

Interpretation and BIRADS.

Help to radiology.

• Proceso

  @startuml
  start
  #yellow: Entendimiento de los procesos hospitalarios;
  #yellow: Entendimiento, selección y preparación de mamografías y metadatos;
  : Preparación de los datos;
  : Modelado y entrenamiento de IA (redes neuronales convolucionales profundas);
  : Evaluación del modelo;
  #yellow: Despliegue del modelo sobre DICOM Web (KPACS a OHIF);
  stop
  @enduml

Server Hospital

ConQuest DICOM server 1.5.0b HGESRVRRX1

Date of this release 20201101

University of California at David (Personal PACS)

Delphi TCP/IP connection

Francios Piette

Lua scripting

sqlite database

UPACS NT PACS System

DCM Standard

8 BSD

Constraints

Web system.

Image Format.

Innovation.

Secretary of Health.=

What are the operating system and another operating system we have to deploy? Windows.

Can we use whatever technology we like? Yes.

What are the conditions we must consider to?

What are system resources?

Will the system access people’s hospitals or hospitals’ LANs?

Are there some user interface guidelines? and Where can I consult for? Yes.

Every patient has n studies and every study has n images.

Every patient

Distributed computing

Everything is ok with cloud platforms except when the time to pay come to.

Preprocessing

Ray

https://docs.ray.io/en/latest/data/pipelining-compute.html#pipelining-datasets

https://docs.ray.io/en/latest/data/api/doc/ray.data.read_images.html#ray.data.read_images

https://docs.ray.io/en/latest/data/examples/ocr_example.html

Download images

cat index.txt | parallel -j+0 "wget -r {}"

Dataset description

Exploratory data analysis

Prediction algorithms

Preprocessing images

TwoViewDensityNet: Two-View Mammographic Breast Density Classification Based on Deep Convolutional Neural Network Mariam Busaleh 1 , Muhammad Hussain 1,* , Hatim A. Aboalsamh 1 , Fazal-e-Amin 2 and Sarah A. Al Sultan 3

ljpeg for Python 3

https://github.com/sanchezcarlosjr/ljpeg

System

PACS over the web

DICOM image viewer

DICOMweb

https://otechimg.com/publications/pdf/dicomweb_white_paper.pdf

https://www.dicomstandard.org/using/dicomweb

• Code

  @startuml
  ExpendienteElectronicoWeb -> EstudiosDeRayosX: iframe /?id=1.2.4.2807152 
  EstudiosDeRayosX --> EstudiosDeRayosXServer: /1.2.4.2807152
  EstudiosDeRayosXServer --> HospitalServerFTP: read directory 1.2.4.2807152 and subdirectories
  EstudiosDeRayosXServer --> DB: ToDo
  EstudiosDeRayosXServer --> EstudiosDeRayosX: {images: [], ...}
  EstudiosDeRayosX --> EstudiosDeRayosX: visualize thumbnails
  EstudiosDeRayosX --> EstudiosDeRayosXServer: visualize /:directory/:subdirectory  example /1.2.4.2807152/2807152.1
  EstudiosDeRayosXServer --> EstudiosDeRayosX: image
  @enduml

Results

Conclusions and limitations

References

https://github.com/OHIF/Viewers

https://pubs.rsna.org/doi/10.1148/radiol.211105

https://github.com/sanchezcarlosjr/Breast-Cancer-risk-estimation-system

https://siim.org/

BI-RADS Terminology for Mammography Reports: What Residents Need to Know. (2023, May 10). Retrieved from https://pubs.rsna.org/do/10.1148/rg.2019180068.pres/full

https://github.com/sanchezcarlosjr/breast-cancer-pipeline

https://github.com/fjeg/ddsm_tools/blob/master/ddsm_tools/ddsm_util.py

https://github.com/fjeg/ddsm_tools/blob/89074b444c54972207f1601fc706939ba9735151/ddsm_tools/ddsm_classes.py#L190