DETAILED ACTION
Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Status of Claims
This action is in response to the reply received 12/26/2025.
Claims 1-2, 4-7, 9-10 were amended 12/26/2025.
Claims 3 and 8 were cancelled 12/26/2025.
Claims 1-2, 4-7 and 9-10 are currently pending and have been examined.
Claim Rejections - 35 USC § 101
35 U.S.C. 101 reads as follows:
Whoever invents or discovers any new and useful process, machine, manufacture, or composition of matter, or any new and useful improvement thereof, may obtain a patent therefor, subject to the conditions and requirements of this title.
35 U.S.C. 101 reads as follows:
Whoever invents or discovers any new and useful process, machine, manufacture, or composition of matter, or any new and useful improvement thereof, may obtain a patent therefor, subject to the conditions and requirements of this title.
Claims 1-2, 4-7 and 9-10 are rejected under 35 U.S.C. 101 because the claimed invention is directed to a judicial exception (i.e., a law of nature, a natural phenomenon, or an abstract idea) without significantly more.
Claims 1-2, 4-7 and 9-10 are drawn to a method and a system which are statutory categories of invention (Step 1: YES).
Independent claims 1 and 6 recite: processing histopathological image data to generate a metastasis risk score in post radical prostatectomy of a patient using, identifying, by a tumour identification and patch generation unit at least one tumour region in at least one input histopathological image corresponding to the histopathological image data, generating, by the tumour identification and patch generation unit a plurality of patches of a pre-defined size from the at least one identified tumour region, wherein generating the plurality of patches comprises; performing patch generation of dense prediction tumour masks for each pixel of the at least one input histopathological image, wherein the patch generation comprises extracting at least one patch having at least 10% tumour from the identified tumour regions; and producing top 2 x N x N patches with the highest tumour percentage; performing, by an image compression unit image compression on each patch of the top 2 x N x N patches to generate compressed histopathological image data by reducing dimensionality of the histopathological image data; performing, by a classification unit input data comprising the compressed histopathological image data and clinical data, wherein the classification is based on generation of concatenated feature vectors during; and generating based on the classification of the input data, wherein represents a risk of metastasis in post radical prostatectomy of the patient.
The recited limitations, as drafted, under their broadest reasonable interpretation, cover certain methods of organizing human activity between a user of the system and a patient, as reflected in the specification, which states that “The embodiments herein disclose employ the trained deep learning system utilizing multimodal data, including histopathology images and clinical data from a group of subjects with long-term followup, for predicting the risk of metastases post radical prostatectomy.” (see: specification paragraph 46). If a claim limitation, under its broadest reasonable interpretation, covers managing personal behavior or relationships or interactions between people, then it falls within the “Certain Methods of Organizing Human Activity” grouping of abstract ideas. The present claims cover certain methods of organizing human activity because they address “Clinical management is complicated by the wide range of disease manifestation and the current method relies on risk-stratifying patients for discovering and treating the most severe lesions.” (see: specification paragraph 3). This problem is addressed “The deep learning system can learn from enormous quantities of data across various modalities and predict the likelihood of metastases in patients post radical prostatectomy.” (see: specification paragraph 21). Accordingly, the claims recite an abstract idea(s) (Step 2A Prong One: YES).
Further, the recited limitations, as drafted, under the broadest reasonable interpretation, cover mathematical relationships by calculating feature vectors from image data. If a claim limitation, under its broadest reasonable interpretation, covers mathematical relationships or mathematical calculations, then it falls within the “Mathematical Concepts” grouping of abstract ideas. Accordingly, the claims recite an abstract idea (Step 2A Prong One: YES).
The judicial exception is not integrated into a practical application. The claims are abstract but for the inclusion of the additional elements including “semantic segmentation network, “an AI score”, and “concatenated feature based system”, “computing device”, “data repository”, “training of the classification unit”, and “wherein classification further comprises selecting N x N features at random from the top 2 x N x N patches generated during the image compression, at each training cycle” are recited at a high level of generality (e.g., that the generating, classifying and displaying is performed using generic computer components designated as various software units running on a generic processor (paragraph 22 of specification) with generic machine learning instructions of a generic neural network (residual neural network (ResNet), paragraph 37) are executed to perform the claimed limitations). Such that they amount to no more than mere instructions to apply the exception using generic computer components. See: MPEP 2106.05(f).
Hence, the additional elements do not integrate the abstract idea into a practical application because they do not impose any meaningful limits on practicing the abstract idea. Accordingly, the claims are directed to an abstract idea (Step 2A Prong Two: NO).
The claims do not include additional elements that are sufficient to amount to significantly more than the judicial exception. As discussed above with respect to integration of the abstract idea into a practical application, using the additional elements to perform the abstract idea amounts to no more than mere instructions to apply the exception using generic components. Mere instructions to apply an exception using a generic component cannot provide an inventive concept. See MPEP 2106.05(f).
Further, the claimed additional elements, identified above, are not sufficient to amount to significantly more than the judicial exception because they are generic components that are configured to perform well-understood, routine, and conventional activities previously known to the industry. See MPEP 2106.05(d). Said additional elements are recited at a high level of generality and provide conventional functions that do not add meaningful limits to practicing the abstract idea. The originally filed specification supports this conclusion at Figure 1, Figure 2 and
Paragraph 21, where “The embodiments herein provides methods and systems for predicting the risk of metastasis in prostate cancer patients. Artificial Intelligence (Al), particularly deep learning-based approaches, employs imaging and categorical clinicopathological data with labels such as "metastasis" or "no metastasis" to train optimized algorithms. AI based approaches can learn from enormous quantities of data across various modalities, such as imaging, molecular markers, and clinicopathologic factors. In contrast to genomic biomarkers, AI systems that exploit digitized images are more affordable and scalable. The embodiments herein employ a multi-modality deep learning system that takes in data from various modalities, such as imaging, molecular markers, and clinicopathologic factors.”
Paragraphs 22 and 42 where “The computing device 102 comprises a tumour identification and patch generation unit 106, an image compression unit 108, and a classification unit 110. Each of the units are executed by a processing unit (not shown in FIG. 1)…The classification unit 110 generates an AI score to determine the risk of metastasis.”
Paragraph 24, where “The identification of the tumour is performed by a semantic segmentation network 204, such as, but not limited to, a transformer-architecture based SegFormer. It is within the scope of the embodiments herein, to employ other semantic segmentation networks, such as a U-Net.”
Paragraph 22, where “FIG. 1 illustrates a concatenated feature based classification system 100 for predicting risk of metastasis in subjects post radical prostatectomy, according to the embodiments herein. The system 100 comprises a computing device 102 coupled to a data repository 104. The data repository 104 comprises clinical data, radiology data, molecular markers, and histopathological images. Examples of the data repository 104 can be, but not limited to, a database, cloud storage, memory, and the like. The data repository 104 and the computing device 102 are connected through at least one wireless connection, or a wired connection over a communication network. In an example, the computing device can be located remotely with respect to the data repository 104.”
Paragraph 32, “In an example herein, the image compression unit 108 generates 2 x N x N patches 316. The image compression unit 108 encodes and compresses each patch of the top 2 x N x N patches with the highest tumour percentage separately and generates an 8 x 8 x 2048 feature block 318. The image compression unit 108 stores the 8 x 8 x 2048 feature blocks 318 in memory, or on disk, or a storage unit (not shown). The image compression step is not trained for metastasis classification. ResNeXt50 SSL weights are utilized primarily for patch-level feature extraction. The image compression unit 108 reduces the dimension of the image from 256x256x3 to 8x8x2048. The initial image passed is of the size 256x256x3 but the output has dimension of 8x8x2048 which is smaller than the original size. The image compression unit 108 removes irrelevant information and reduces the size of the image.”
Paragraph 36, “The classification unit 110 employs a classifier 420. In an embodiment herein, the classifier 420 can use a deep learning model, such as, ResNetl8. It is within the scope of the embodiments herein to use other deep learning models such as, but not limited to, convolutional neural networks (CNNs) and their variants, such as 3D-CNNs or residual networks (ResNets) for the classification. ResNetl8 is a deep neural network architecture that belongs to the ResNet family of models. The ResNetl8 architecture comprises of 18 layers, including a convolutional layer, a max-pooling layer, and several residual blocks. The residual blocks are the key feature of ResNetl8, and they allow the model to learn increasingly complex features from the input images. Each residual block in ResNetl 8 comprises of two convolutional layers and a shortcut connection that bypasses the convolutional layers. The shortcut connection allows the model to learn from the identity mapping of the input data and helps to mitigate the vanishing gradient problem that can occur in deep neural networks. The convolutional layers within each residual block use 3x3 filters, and the number of filters is increased gradually throughout the layers to allow the model to learn increasingly complex features. The last layer of ResNetl 8 is a fully connected layer that performs the final classification of the input image. ResNetl 8 is advantageous because of its relatively small size, which makes it computationally efficient and easy to train, while still achieving high accuracy on many tasks.”
Paragraph 37, “The classification unit 110 receives two inputs: one is the clinical data and the other one is the image data. In an example, as explained above, the ResNetl8, which is the CNN part of the classification unit 110, processes the image data. A two-layer fully connected network, also known as the 2 layer MLP, processes the clinical data which is passed to it in vector form. In an example, the vector can be a 256-dimensional vector 422. The classification unit 110 concatenates this vector with ResNetl8's penultimate layer output and the concatenated network is passed through a final layer. The output from the final layer is passed to softmax function to obtain an AI risk score for determining the risk of the metastasis.”
Paragraph 34, “at each training cycle, the classification unit 110 selects N x N features at random from the 2 x N x N patch representation generated during image compression. This random selection of training patches is an effective strategy for enhancing performance on tests.”
Viewing the limitations as an ordered combination, the claims simply instruct the additional elements to implement the concept described above in the identification of abstract idea with route, conventional activity specified at a high level of generality in a particular technological environment.
Hence, the claims as a whole, considering the additional elements individually and as an ordered combination, do not amount to significantly more than the abstract idea (Step 2B: NO).
Dependent claims 2, 4-5 and 7, 9-10 when analyzed as a whole, considering the additional elements individually and/or as an ordered combination, are held to be patent ineligible under 35 U.S.C. 101 because the additional recited limitations fail to establish that the claims are directed to an abstract idea without significantly more. Claim 2 and 7 recite transmitting, receiving, and scheduling healthcare data from the software units run on the generically recited computing device using generic machine learning algorithms as shown in the parent claims above.
Claims 4 and 9 further recite “pre-processing the top 2 x N x N patches with the highest tumour percentage using a Diversity Inducing Non-parametric Optimization (DINO) approach, wherein the pre-processing normalizes size and intensity of values of the at least one patch”, which are nominal or tangential addition to the abstract idea and amount to insignificant post-solution activity concerning an insignificant application. The pre-processing is described as an optimization technique used for training neural networks (paragraph 30 of the specification). The addition of an insignificant extra-solution activity limitation does not impose meaningful limits on the claim such that is it not nominally or tangentially related to the invention. In the claimed context, these claimed additional elements are incidental to the performance of correlating healthcare data as outlined in the recitations above. See: MPEP 2106.05(g).
Claims 5 and 10 further recite “creating a two-layer fully connected network” which is recited at a high level of generality (e.g., that the creating of layers in a machine learning network are performed using generic computer components with instructions are executed to perform the claimed limitations) as shown in the specification paragraph 37. Such that they amount to no more than mere instructions to apply the exception using generic computer components. See: MPEP 2106.05(f).
These claims fail to remedy the deficiencies of their parent claims above, and therefore rejected for at least the same rationale as applied to their parent claims above, and incorporated herein.
Claim Rejections - 35 USC § 103
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
Claim(s) 1-2, 4-7 and 9-10 is/are rejected under 35 U.S.C. 103 as being unpatentable over Esteva (WO 2023/0107297 A1) in view of Courtiol (US 2021/0271847 A1).
CLAIM 1-
Esteva teaches the limitations of:
A method for processing histopathological image data to generate a metastasis risk score in post radical prostatectomy of a patient, (Esteva teaches predicting risk of prostate cancer outcomes using machine learning including patients with risk of outcome after providing a treatment such as a prostatectomy and outputs a risk score for each patient (para [0060, 00123, 00192])) the method comprising: identifying, by a tumour identification and patch generation unit at least one tumour region in at least one input histopathological image corresponding to the histopathological image data (Esteva teaches identifying the tumor using patch generation that feeds patches through the model to visualize the patches of tissue to identify tumors and is used to train the data by inputting a histopathological image and histopathological slides of a patient during their treatment to determine efficacy, the treatment can include a prostatectomy (para [0017, 0020, 00132, 00123, 00189]))
generating, by the tumour identification and patch generation unit a plurality of patches of a pre-defined size from the at least one identified tumour region, wherein generating the plurality of patches comprises; (Esteva teaches that the tumor identification process includes generating a patch with a predetermined size from the region of interest which can include a tumor’s region of interest and Esteva teaches identifying the tumor using patch generation that feeds patches through the model to visualize the patches of tissue to identify tumors and is used to train the data by inputting a histopathological image of a patient during their treatment to determine efficacy, the treatment can include a prostatectomy (para [0017, 0020, 00132, 00123, 00159, 0075, 0065, Figure 11)) (]))
performing, by an image compression unit, image compression each patch of the top 2 x N x N patches to generate compressed histopathological image data by reducing dimensionality of the histopathological image data; (Esteva teaches that the image processing procedure may compress an image by compressing a massive image quilt into a compact representation using H X W X 128 wherein H could be 2 and N can be 128 in order to compress dimensionality of the histopathological feature images (i.e., reducing dimensionality by reducing the size of the image as described in the specification paragraph 32) and is optimized (i.e., top) (para [0075, 00159,0196]))
performing, by a classification unit classification of input data comprising the compressed histopathological image data and clinical data, wherein the classification is based on generation of concatenated feature vectors during training of the classification unit, (Esteva teaches using risk classification from input image data to predict risk of metastasis in the patients during their treatment to determine efficacy, the treatment can include a prostatectomy. Esteva further teaches that the classification is based on training concatenated feature vectors from the compressed data for classifiers (i.e., classification unit) (para [00132, 00123, 00216, 00174, 00199, 00192, 0196, 00159]))
and generating an AI score based on the classification of the input data, wherein the AI score represents a risk of metastasis in post radical prostatectomy of the patient (Esteva teaches that based on the neural network calculations, a risk stratification score tis attached from highest to lowest risk of metasis and the data can include patients that during their treatment received a prostatectomy (para [00132, 00123, 00192, 00216, 0116], Figure 24))
Esteva does not explicitly teach, however Courtiol teaches:
using a concatenated feature based system (Courtiol teaches a system that uses ResNet-50 neural network outputting vectors of dimensional 2048 which is what the claimed concatenated feature based system does in the specification paragraph 41, paragraph 37 under broadest reasonable interpretation (para [0055-056]))
using a semantic segmentation network; (Courtiol teaches using a semantic segmentation neural network to classify tumor images (para [0031, 0035, 0046]))
performing path generation of dense prediction tumour masks for each pixel of the at least one input histopathological image, wherein the patch generation comprises extracting at least one patch having at least 10% tumour from the identified tumour regions (Courtiol teaches using the dense layers to generate a feature set to predict the tumor size based on the input histopathological images and the identified regions have at least a percentage set by the user interface which may include ten percent and includes classifying each pixel of the input image (para [0034, 0038, 0060, 0067, 0083-84, 0080, 0062, claim 2]))
and producing top 2 x N x N patches with the highest tumour percentage; (Courtiol teaches that the N tiles are output into two vectors to get a set of global labels for the image including tumor size and can show the rankings of the top N tile scores that include highest tumor size (i.e., tumor percentage) (para [0080, 0082, 0078]))
wherein the classification further comprises selecting N x N features at random from the top 2 x N x N patches generated during the image compression, at each training cycle; (Courtiol teaches that the N tiles are output into two vectors and can show the rankings of the top N tile scores that include highest tumor size (i.e., tumor percentage) that are generated by reducing dimensionality (i.e., compressing the image). Courtiol further teaches that the classification system randomly samples the tiles to reduce the number of times used in the neural network computations (i.e., at each training cycle, the tiles are reduced) (para [0039, 0080, 0082, 0078, 0056, 0067]))
It would have been obvious to one of ordinary skill in the art at the time the invention was filed to modify the tumor identification system of Esteva to integrate the application of using a semantic segmentation network to identify tumors of Courtiol with the motivation of improving patient outcomes by assisting pathologists to identify previously unknown structures and regions of interest (see: Courtiol, paragraph 2).
CLAIM 2-
Esteva in view of Courtiol teaches the limitations of claim 1. Regarding claim 2, Esteva further teaches:
wherein identifying, by a tumour identification and patch generation unit, tumour regions in the at least one input histopathological images, comprises: performing the tumour identification by first encoding the at least one input histopathological images (Esteva teaches identifying the tumor using patch generation that feeds patches through the model to visualize the patches of tissue to identify tumors and is used to train the data by inputting a histopathological image of a patient during their treatment to determine efficacy, the treatment can include a prostatectomy (para [0017, 0020, 00132, 00123]))
using a set of convolutional layers to obtain a set of feature maps; (Esteva teaches using the convolutional layers to obtain a set of filters to detect a specific type of feature in an activation map (i.e., feature maps) (para [0086, 0088, 0089]))
processing the set of feature maps by a series of transformer blocks to extract high-level representations that capture both local and global context; (Esteva teaches using attention mechanisms (i.e., transformers) to define the most relevant information (i.e., extracting high-level representations) in order to define where the more relevant information in the input sequence is located (i.e., mapping features locally) and can use pooling layers that comprise global pooling layers to combine outputs (i.e., globally mapping data) (para [0089, 0092-0093]))
Esteva does not explicitly teach, however Courtiol teaches:
and transmitting an output of the series of transformer blocks to a decoder network that produces a dense prediction tumour masks for each pixel of the at least one input histopathological images (Courtiol teaches using image score vectors as input to a dense multilayer neural network (i.e., feature sets are transform into features of the layers of a network that decodes data) to produce a classifier of the tumor for each pixel of the histopathological images (para [0060, 0067, 0083-84, 0080]))
It would have been obvious to one of ordinary skill in the art at the time the invention was filed to modify the tumor identification system of Esteva to integrate the application of using a semantic segmentation network to identify tumors of Courtiol with the motivation of improving patient outcomes by assisting pathologists to identify previously unknown structures and regions of interest (see: Courtiol, paragraph 2).
CLAIM 4-
Esteva in view of Courtiol teaches the limitations of claim 1. Regarding claim 4, Esteva does not explicitly teach, however Courtiol teaches:
wherein performing, by the image compression unit, the image compression on the at least one patch comprises: pre-processing the top 2 x N x N patches with the highest tumour percentage (Courtiol teaches that the N tiles are output into two vectors to get a set of global labels for the image including tumor size that can be tiled into image subsets (i.e., compressed) and can show the rankings of the top N tile scores that include highest tumor size (i.e., tumor percentage) (para [0039, 0080, 0082, 0078])) by using a Diversity Inducing Non-parametric Optimization (DINO) approach, wherein the pre-processing normalizes size and intensity of values of the at least one patch; (Courtiol teaches optimizing the model based on the differences of the weights within a threshold to encode images into feature vector sets to normalize size and intensity of values (which reads, under broadest reasonable interpretation, on the specification recitation of what a DINO approach is paragraph 30) (para [0034, 0073], claim 4, claim 5))
and encoding and compressing each of the patch of the top 2 x N x N patches with the highest tumour percentage separately and generating an 8 x 8 x 2048 feature block (Courtiol teaches the autoencoding and compressing of the extracted feature vectors to reduce the dimensionality of the features to rank the highest scores which can be defined by percentage (i.e., the tumor image is classified with the highest tumor percentage) and Figure 5 shows how the 2 N x N tiles are generating into 8 x 8 x 2048 feature block (para [0034, 0056, 0058, 0060, 0062], Figure 5))
It would have been obvious to one of ordinary skill in the art at the time the invention was filed to modify the tumor identification system of Esteva to integrate the application of using a semantic segmentation network to identify tumors of Courtiol with the motivation of improving patient outcomes by assisting pathologists to identify previously unknown structures and regions of interest (see: Courtiol, paragraph 2).
CLAIM 5-
Esteva in view of Courtiol teaches the limitations of claim 1. Regarding claim 5, Esteva further teaches:
wherein classification of the input data by the classification unit, for predicting the risk of metastasis in subjects post radical prostatectomy, comprises: (Esteva teaches predicting risk of prostate cancer outcomes using machine learning including patients with risk of outcome after providing a treatment such as a prostatectomy by classifying images (para [0060, 00123], Figure 11))
creating a two-layer fully connected network, using an input data and generating a 256- dimensional vector from the input data; (Esteva teaches using a 2 layer grid (H x W) by using input data and generating a 256 dimensional patches and can use the machine learning algorithm to create fully-connected layers (para [00159, 0086-0088]))
performing metastasis classification by transmitting the concatenated vector through the two-layer fully connected network; (Esteva teaches performing classification of the images to determine metastasis by transmitting a concatenation of the image vector to be processed by the neural network that includes the multiple fully-connected layers (para [0199, 00159, 0086-0088, 00225]))
Esteva does not explicitly teach, however Courtiol teaches:
forming a tensor of size 8N x 8N x 2048 by concatenating patch representations of the N x N features; (Courtiol teaches that the features are concatenating to sort the tiles and Figure 5 shows how the 2 N x N tiles are generating into 8 x 8 x 2048 feature block (para [0034, 0056, 0058, 0060, 0062, 0085], Figure 5))
concatenating the 256-dimensional vector with the 8N x 8N x 2048 tensor; (Courtiol teaches that the features are concatenating to sort the tiles and Figure 5 shows how the 2 N x N tiles are generating into 8 x 8 x 2048 feature block and received 256 features for each feature vector (para [0034, 0056, 0058, 0060, 0062, 0085, 0082], Figure 5))
It would have been obvious to one of ordinary skill in the art at the time the invention was filed to modify the tumor identification system of Esteva to integrate the application of using a semantic segmentation network to identify tumors of Courtiol with the motivation of improving patient outcomes by assisting pathologists to identify previously unknown structures and regions of interest (see: Courtiol, paragraph 2).
CLAIM 6-
Claim 6 is significantly similar to claim 1 and is rejected upon the same prior art as claim 1.
CLAIMS 7, 9-10-
Claims 7, 9-10 are significantly similar to claims 2, 4-5 and are rejected upon the same prior art as claims 2, 4-5 respectively.
Response to Arguments
The arguments filed 12/26/2025 have been fully considered.
The arguments pertaining to the Claim Objections are persuasive. The claims have been sufficiently amended and the objections are withdrawn.
The arguments pertaining to the 101 rejection are not persuasive. Applicant argues that the claims are directed to a technological improvement in computer-based medical image processing and improves the functioning of the computer itself. Examiner respectfully disagrees. The functions argued are representative of the abstract idea. The claims here are not directed to a specific improvement to computer functionality that amount to a practical application. Rather, they are directed to the use of conventional or generic technology in a well-known environment, without any claim that the invention reflects an inventive solution to a technical problem presented by combining the two. In the present case, the claims fail to recite any elements that individually or as an ordered combination transform the identified abstract idea(s) in the rejection into a patent-eligible application of that idea
Further, not every claim that recites concrete, tangible components escapes the reach of the abstract-idea inquiry. (See, e.g., Alice, 134). It is well-settled that mere recitation of concrete, tangible components that are generic is insufficient to confer patent eligibility to an otherwise abstract idea. In order to amount to an inventive concept, the components must involve more than performance of “’well-understood, routine, conventional activities’ previously known to the industry.” (Alice, 134 S. Ct. at 2359 (quoting Mayo, 132 S.Ct. at 1294)). The originally filed specification was investigated and found to support this conclusion.
Applicant further argues that the claimed invention is similar to that of Enfish. Examiner respectfully disagrees, as there is no improvement on technology. The machine learning described in the claimed invention that uses layers and concatenated feature vectors to process the image data is the well-known ResNet machine learning. ResNet (residual neural network) is well-known in the art as a generic machine learning model and functions as a generic machine learning network as shown in the specification paragraphs 36-38. The performance is enhanced on the output of data not on the computer functionality itself.
Applicant further argues that the amendments that the technical solution of generation of the AI score representing risk of metastasis in post radical prostatectomy of the patient is a practical application. Examiner respectfully disagrees as improving diagnostics the medical field by improving data output to a healthcare practitioner and patient as evidenced in the specification show that the claims are directed to organizing human activity. The “medical field” is not necessarily a “technical field”, nor is a treatment effected. Classen is an example of adding a meaningful limitation to the claims that create a practical application, however Classen integrated the results of the analysis into a specific and tangible method that resulted in the method “moving from abstract scientific principle to specific application” (Classen Immunotherapies Inc. v. Biogen IDEC). The current claimed limitations fail to provide this practical application and the 101 rejection is maintained.
The arguments pertaining to the 103 rejection are not persuasive. Applicant argues that the limitations as Esteva teaches extracting features from histopathological images using convolutional neural networks and correlating such features with clinical outcomes at a slide or tile level without tumour-precentage-based patch ranking, randomized training-cycle feature selection or concatenated feature-based classification. Examiner respectfully disagrees. The current claimed invention does not explicitly claim tumour-percentage-based patch ranking or randomized training-cycle feature selection. Under broadest reasonable interpretation, Esteva teaches using patch generation of dense layers (i.e., masks) for pixels of the images to extract tumor image data from identified tumor regions as shown above. Esteva further teaches compressing histological image data by reducing dimensionality of the image data in order to optimize patch selection as shown above. These features are what is claimed in the current claimed invention, not what is being argued. Selecting N x N features at randrom the top x N x N patches is taught by Courtiol, not Esteva.
Applicant further argues that Courtiol uses a weakly supervised framing network. Examiner respectfully disagrees. Couritol teaches using the same supervised learning framework as the claimed invention (ResNet) as shown above. Courtiol further teaches that the classification system randomly samples the tiles to reduce the number of times used in the neural network computations (i.e., random selection of features during training to form the concatenated feature vectors).
Under broadest reasonable interpretation, the references of Esteva and Courtiol teach the claimed invention and follow proper USPTO guidelines in their combination.
The dependent claims rely on the arguments of the independent claims and are rejected for the reasons stated above.
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure. Varkuti (WO 2016116449 A1) teaches using vector tumor growth masks to determine the percentage of tumor growth similar to the current claimed invention.
Applicant's amendment necessitated the new ground(s) of rejection presented in this Office action. Accordingly, THIS ACTION IS MADE FINAL. See MPEP § 706.07(a). Applicant is reminded of the extension of time policy as set forth in 37 CFR 1.136(a).
A shortened statutory period for reply to this final action is set to expire THREE MONTHS from the mailing date of this action. In the event a first reply is filed within TWO MONTHS of the mailing date of this final action and the advisory action is not mailed until after the end of the THREE-MONTH shortened statutory period, then the shortened statutory period will expire on the date the advisory action is mailed, and any nonprovisional extension fee (37 CFR 1.17(a)) pursuant to 37 CFR 1.136(a) will be calculated from the mailing date of the advisory action. In no event, however, will the statutory period for reply expire later than SIX MONTHS from the mailing date of this final action.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to KIMBERLY A SASS whose telephone number is (571)272-4774. The examiner can normally be reached 7AM-5PM (EST).
Examiner interviews are available via telephone, in-person, and video conferencing using a USPTO supplied web-based collaboration tool. To schedule an interview, applicant is encouraged to use the USPTO Automated Interview Request (AIR) at http://www.uspto.gov/interviewpractice.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, JASON DUNHAM can be reached at 571-272-8109. The fax phone number for the organization where this application or proceeding is assigned is 571-273-8300.
Information regarding the status of published or unpublished applications may be obtained from Patent Center. Unpublished application information in Patent Center is available to registered users. To file and manage patent submissions in Patent Center, visit: https://patentcenter.uspto.gov. Visit https://www.uspto.gov/patents/apply/patent-center for more information about Patent Center and https://www.uspto.gov/patents/docx for information about filing in DOCX format. For additional questions, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
/K.A.S./ Examiner, Art Unit 3686
/JASON B DUNHAM/ Supervisory Patent Examiner, Art Unit 3686
Prosecution Insights