Image analysis system for identifying lung features
Methods and apparatuses for identifying lung features are provided herein.
1 . A method of determining segment boundaries of lung segments, the method comprising:
receiving image data forming a three-dimensional representation of at least a part of a lung, wherein the lung is divided into multiple lung lobes, wherein each lobe is divided into multiple lung segments;
computationally identifying, using the image data, at least one anatomical feature within the lung, wherein the at least one anatomical feature comprises one or more of the following: a network of arteries, a network of veins, a network of airways, some or all of lung parenchyma, and one or more fissures of the lung;
computationally identifying, using the at least one anatomical feature, all segment boundaries of at least one lung segment within the lung, wherein the segment boundaries comprise surfaces between the at least one lung segment and adjacent lung segments; and
generating a representation of all or part of the lung showing the segment boundaries of the at least one lung segment within the lung.
2 . The method of claim 1 , wherein the image data comprises a one or both of a CT scan and an MRI.
3 . The method of claim 1 , wherein the at least one anatomical feature comprises some or all of the lung parenchyma and the network of arteries.
4 . The method of claim 1 , wherein the at least one anatomical feature comprises some or all of the lung parenchyma and the network of veins.
5 . The method of claim 1 , wherein the at least one anatomical feature comprises some or all of the lung parenchyma and the network of airways.
6 . The method of claim 1 , wherein the at least one anatomical feature comprises the one or more fissures of the lung.
7 . The method of claim 1 , wherein computationally identifying all segment boundaries of the at least one lung segment comprises:
computationally identifying, using the at least one anatomical feature, one or more of the following: segmental branches of arteries and related subtrees, segmental branches of veins, segmental branches of airways and related subtrees, and the fissures of the lung;
generating segment boundaries between adjacent lung segments in the lung through analysis of segmental branches of arteries and related subtrees, segmental branches of veins, segmental branches of airways and related subtrees, and/or fissures of the lung; and
combining the segment boundaries between adjacent lung segments to define all of the segment boundaries of the at least one lung segment within the lung.
8 . The method of claim 1 , further comprising:
computationally identifying, using the at least one anatomical feature, lobe boundaries for a plurality of lobes in the lung.
9 . The method of claim 1 , further comprising:
computationally identifying, using the image data, a lesion within the lung; and
computationally determining, from the image data, that the lesion is located in a given lung segment of the at least one lung segment or in a region spanning two or more lung segments.
10 . The method of claim 9 , further comprising:
computationally measuring, from the image data, a minimum distance between the lesion and a segment boundary of any of the lung segments.
11 . A method of identifying segment boundaries of lung segments, the method comprising:
receiving image data forming a volumetric representation of at least a part of a lung, wherein the lung is divided into multiple lobes, wherein each lobe is divided into multiple lung segments;
computationally identifying, using the image data, a network of arteries;
computationally identifying, using the image data, a network of bronchi;
computationally identifying, using the image data, a network of veins; and
computationally identifying, based on the identified network of arteries, network of bronchi, and network of veins, segment boundaries of a plurality of lung segments within at least one lobe of the lung, wherein the segment boundaries comprise surfaces between adjacent lung segments.
12 . The method of claim 11 , wherein the image data comprises one or both of a CT scan and an MRI.
13 . The method of claim 11 , wherein computationally identifying the network of arteries comprises:
computationally identifying a tube-like structure in the image data, wherein the tube-like structure is identified by identifying a set of gradient changes within the image data; and
computationally determining, using the image data and based on how the tube-like structure branches within the lung, that the tube-like structure is part of the network of arteries.
14 . The method of claim 11 , wherein computationally identifying the network of bronchi comprises:
computationally identifying a tube-like structure in the image data, wherein the additional tube-like structure is identified by identifying a set of gradient changes within the image data; and
computationally determining, using the image data and based on how the additional tube-like structure branches within the lung, that the tube-like structure is part of the network of bronchi.
15 . The method of claim 11 , wherein computationally identifying the network of veins comprises:
computationally identifying a tube-like structure in the image data, wherein the tube-like structure is identified by identifying a set of gradient changes within the image data; and
computationally determining, using the image data and based on how the tube-like structure branches within the lung, that the tube-like structure is part of the network of veins.
16 . The method of claim 11 , wherein computationally identifying the segment boundaries of the plurality of lung segments comprises:
computationally identifying a volume within the image data that receives blood from a section of the network of arteries.
17 . The method of claim 11 , wherein computationally identifying the segment boundaries of the plurality of lung segments comprises:
computationally identifying a volume within the image data that receives air from a section of the network of bronchi.
18 . The method of claim 11 , further comprising:
computationally identifying, using the image data, one or more intersegmental veins; and
computationally refining, based on the one or more intersegmental veins, the segment boundaries of the plurality of lung segments of the at least one lobe.
19 . A method of determining segment boundaries of lung segments, the method comprising:
receiving image data forming a three-dimensional representation of at least a part of a lung, wherein the lung is divided into multiple lobes, wherein each lobe is divided into multiple lung segments;
computationally identifying, using the image data, one or more intersegmental veins within the lung;
computationally identifying segment boundaries of at least one lung segment within the lung based at least in part on the identified one or more intersegmental veins, wherein the segment boundaries comprise surfaces between the at least one lung segment and adjacent lung segments; and
generating a representation of all or part of the lung showing the segment boundaries of the at least one lung segment within the lung.
20 . A method of determining segment boundaries of lung segments, the method comprising:
receiving image data forming a three-dimensional representation of at least a part of a lung, wherein the lung is divided into multiple lobes, wherein each lobe is divided into multiple lung segments;
computationally identifying, using the image data, one or more fissures of the lung;
computationally identifying segment boundaries of at least one lung segment within the lung based at least in part on the identified one or more fissures, wherein the segment boundaries comprise surfaces between the at least one lung segment and adjacent lung segments; and
generating a representation of all or part of the lung showing the segment boundaries of the at least one lung segment within the lung.