Patent Profile: Patent Issued for Method of Differentiating Pancreatic Cancer from Benign Pancreatic Disease

By Josh Bosman --

The U.S. Patent and Trademark Office recently issued U.S. Patent No. 8,632,983, which is entitled "Biomarkers for pancreatic cancer and diagnostic methods."  The '983 patent, which is assigned to the Van Andel Research Institute and the University of Pittsburgh, contains claims to a method for differentiating pancreatic cancer from a benign pancreatic disease.

The invention provides a method for diagnosing pancreatic cancer; more specifically, a method for differentiating pancreatic cancer from a benign pancreatic disease in a patient having or suspected of having a pancreatic disease.  The invention could be particularly helpful in the early diagnosis of pancreatic cancer.  Pancreatic cancer is typically diagnosed at a more advanced stage, and treatment options are limited, leading to five year survival rates of less than 5%.  The inventors indicate that a blood-based diagnostic test for pancreatic cancer would be especially valuable, potentially increasing survival rates and saving resources.

The inventors used antibody arrays to measure the levels of a carbohydrate antigen, CA 19-9, on multiple proteins in 420 individual biological samples from patients with pancreatic adenocarcinoma or pancreatitis.  It is known that the CA 19-9 antigen is elevated in serum from the majority of pancreatic cancer patients, but using CA19-9 in diagnostic tests is discouraged in the guidelines from the American Society of Clinical Oncology due to both false positive and false negative readings.  Surprisingly, the inventors discovered that a subset of cancer patients showed elevations of the CA 19-9 antigen on the mucin proteins MUC1, MUC5AC, or MUC16.  By combining measurements of total CA 19-9 with CA 19-9 on the MUC1, MUC5AC, or MUC16 proteins, the inventors determined that they could differentiate, with sensitivity and specificity, whether the sample was from a patient with pancreatic cancer or a patient with a benign pancreatic disease.

The '983 patent has one independent claim which recites:

1.  A method for differentiating pancreatic cancer from a benign pancreatic disease, comprising the steps:
    obtaining a patient biological sample from a patient having or suspected of having a pancreatic disease;
    assaying the patient biological sample (a) to detect a total level of CA 19-9 antigen in the patient biological sample and (b) to detect a glycan level in MUC16 in the patient biological sample;
    comparing the total level of CA 19-9 antigen in the patient biological sample to a statistically validated threshold for total CA 19-9 antigen, which statistically validated threshold for total CA 19-9 antigen is based on a total level of CA 19-9 antigen in comparable control biological samples from patients having a benign pancreatic disease; and
    comparing the glycan level in the MUC16 in the patient biological sample to a statistically validated threshold for the MUC16, which statistically validated threshold for the MUC16 is based on a glycan level in the MUC16 in comparable control biological samples from patients having a benign pancreatic disease;
    wherein (a) a different level of total CA 19-9 antigen in the patient biological sample as compared to the statistically validated threshold for total CA 19-9 antigen and (b) a different level of glycan level in the MUC16 in the patient biological sample as compared to the statistically validated threshold for the MUC16 indicate that the patient has pancreatic cancer rather than a benign pancreatic disease.

A review of the file history of the '983 patent indicated that a rejection under 35 U.S.C. § 101 was not issued.

Comments

[Diana Cox]

Must be nice to get your application to the right GAU, with the right Examiner.

I handle a lot of diagnostics cases, with claims that look very much like these claims, including statistically validated thresholds, and these claims would not pass muster at all with the Examiners to whom my cases are assigned:

- The assay step: nothing transformative about it; it could be done by a Star Trek tricorder*.

- There is no practical application of the results of this method (nevermind that it is self-evident) - it is merely a statement of a natural law.

(And why do people insist on including a step of obtaining the patient sample? It is rarely done by the clinical lab - include the description of the patient in the assay test: "detecting and quantitating in a sample obtained from a patient having or suspected of having ...")

* PerkinElmer v. Intema:
The purported transformation resulting from “assaying a sample” is insufficient since it could be performed “without transforming the [sample], should science develop a totally different system for [assaying for a biochemical screening marker] that did not involve such a transformation.”

Meanwhile, the publication of the "new" Mayo/Myriad Examination Guidelines has been pushed back once again, to the first week of March, as the Examiners are currently being trained on them. I'm betting they are going to be even more onerous.

[examiner]

Diana:
I am an examiner in one of those AUs and I (mostly) agree with you. But I hope you understand that a lot of us also agree with your sentiments. However, it's largely non-examiners that are responsible for the "guidelines" and we are required to follow them, even if we personally disagree.

That said, the analysis we are taught is this:
1. Natural correlation/law of nature?
2. If yes to 1, then practical application (i.e. not "monopolizing" correlation/law of nature)?
3. If yes to 2, then look at each step INDIVIDUALLY and if done at least once= "routine, conventional, etc."

As to "statistically validated threshold," this is way too broad and encompasses virtually any "threshold" that's been "validated" (i.e. any method to validate a threshold and/or any threshold). There's no definition of this in the claim (and probably a very broad definition in the spec)= not distinguishable over prior art.

Unless the detection of both CA 19-9 AND MUC16 is new (or any other step is new), Prometheus stands in the way. At least according to our guidelines.

Although, from the sound of it, the inventors have some strong secondary evidence...I've always wondered if convincing unexpected results, for example, could overcome a Prometheus rejection based on "routine, conventional" steps. In other words, does unexpected result= not routine and conventional to perform this step with this correlation. But the response I'll likely get is: doesn't matter, step ITSELF= known= "routine, conventional."

That said, APPEAL. Please...

[Will R.]

Examiner: Glad to see that you understand how to do the analysis. You are correct that claims that cover "making new correlations" based on data that was previously gathered in the prior art (regardless of why the data was gathered) are ineligible claims.

This ineligibility is the correct result, for the reasons given by the Supreme Court in Prometheus v. Mayo (among other reasons). If such claims are routinely granted, clinical diagnosis will soon devolve into a cesspool whereby practitioners of the prior art can be dragged into court for allegedly "correlating" some data that was obtained perfectly legally. That's an absolutely repulsive result and it's why Prometheus v. Mayo will never be overturned. Patents can not be used to protect information or scientific facts (or non-facts where the underlying data/analysis is bunk) but that's exactly what an improvidently granted "determine and infer" claim does.

Without looking at the file history, I can't be sure whether the combined data collection steps in the specific claim discussed here are non-obvious. Note that if the data collection steps are non-obvious and transformative (i.e., eligible on their own terms), then there is no need to recite the correlation at all. The method would be both eligible and patentable over the prior art. It's when the applicant attemps to overcome the prior art by reciting the ineligible "new correlation" that the claim runs into problems.

Please let me know if you have any questions about this. If anyone at the USPTO is still confused about this straightforward problem, they need to be educated quickly. Never confuse "utility" with eligibility. They are non-identical concepts. Put another way, there are many, many, many very useful methods that are ineligible for patenting and thank goodness for that.

[Diana Cox]

Even with statistical validation method(s) recited in the claim, the point is, the THRESHOLD level is a value (not even that is present in the claims of the above patent) that is not itself obvious, absent the inventive input of Applicants which makes it ascertainable, and thus useful.

And last time I looked, Section 101 still recites the word "discovery." The correlation is a discovery, as is the threshold and what it means for patient treatment. And the process of conducting the assay is a practical application of the discovery of the correlation.

The requirement to recite, as an ADDITIONAL practical application, how a doctor uses the information about the assay results to inform the prescribing of treatment - taking into account other patient-specific factors that have nothing to do with the laboratory test - is ridiculous (to say nothing of the enforcement issue of the doctor being an independent entity with no agency relationship to the clinical lab). Of course a doctor uses the information - in conjunction with a myriad of other factors (including the patient's own opinions) - to prescribe treatment. The practical application is inherent in the fact that the test was ordered by the doctor in the first place.

[Skeptical]

I am amazed that any professional in the patent arena would type "Never confuse "utility" with eligibility. They are non-identical concepts"

101 is the law of eligibility.
Utility is a 101 concept.

Thus, utility is fully a subset of the law of eligibility.

Now, if this latest sockpuppet of Malcolm Mooney had intended to write "do not conflate utility with the categories of patent eligible subject matter," I would be less offended by the charlatan's latest attempt to obfuscate the law.

[Will R.]

Diane: "Section 101 still recites the word "discovery." The correlation is a discovery, as is the threshold and what it means for patient treatment."

First, 101 does not recite the word "discovery." The relevant text is 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."

Whether you like it or not, the Supreme Court has decided (in a decision that nearly everyone agrees is correct except for a tiny minority of invested individuals) you can't patent a correlation merely by dressing the claim up with old, conventional steps of data collection. This is as it should be and I'm happy to explain why if you are confused about the reasoning. You surely do not need to be reminded that Prometheus' theory of infringement did not depend on any treatment or advice given to patients by Mayo, but simply on the fact that Mayo's doctors possessed legally obtained data and thought about the ineligible correlation that Prometheus dedicated to the public in their patent application (Mayo, of course, had "discovered" their own distinct correlation).

"The practical application is inherent in the fact that the test was ordered by the doctor in the first place."

Again, you are confusing utility with eligibility. There is more to eligibility than a mere utility requirement, thank goodness.

[Will R.]

Skeptic: "I am amazed that any professional in the patent arena would type "Never confuse "utility" with eligibility. They are non-identical concepts""

They are non-identical concepts and your comment just underscores that fact.

There's nothing amazing about you shooting yourself in the foot, of course, as that seems to be a specialty of yours.

Maybe Kevin can help you explain what you were trying to say. Or maybe he can help you learn to say "I'm sorry." I wish him luck.

[Skeptical]

Tell me, "Will R." are you the poster commonly known as Malcolm Mooney, the self same poster who swore he did not use sockpuppets?

Tell me, while you are at, how many sockpuppets have you used here at PatentDocs.

Then tell me why do you engage in such duplicity and hypocrisy? Then tell me if you feel that lying is OK on this blog as well, since you feel that such is perfectly acceptable at your usual haunt.

And I am truly sorry that PatentDocs has this rash of infection of Malcolm Mooneyism.

Are you sorry for what you should be sorry for? Of that, I remain....