25 Surveillance Programs

Emerging Applications: Clinical Case Review

These are applications that are evolving and have the potential to snowball and dramatically expand the applications of prostate MRI.

We are going to start with surveillance programs. Dr. Kora is going to talk about active surveillance, and she will be followed by Dr. Hader.

Multiparametric MRI in Surveillance Programs

Good afternoon. We are going to continue with the multiparametric MRI applications in a little different direction, as Dr. Weiner already mentioned. We are going to talk about the surveillance programs over the next 20 minutes.

I am going to discuss what is active surveillance and what are the current guidelines and what are the results in terms of the risk stratification, survival, quality of life, and dropout rates that are being reported.

We have a big program at Hopkins in active surveillance led by Valentine Carter, and I will share some of the experience from our programs and then conclude with a few cases.

What is Active Surveillance?

The active surveillance is monitoring with intent of cure of cancer and intervene when needed. It is being used as an alternative to immediate intervention to limit the overtreatment of low risk and non-life threatening prostate cancer.

The primary goals of monitoring are to avoid unnecessary treatment of indolent disease, and of course, associated morbidity, and also promptly identify evidence of high risk disease that warrants immediate intervention.

As we know, about 10 to 33% of patients on active surveillance will eventually undergo reclassification. However, the remainder of the patients who are in the program will avoid unnecessary interventions. It is a very important management strategy.

Recommendations for Active Surveillance

The recommendations for active surveillance are very individualized, and several factors are being considered, such as patient's life expectancy, disease characteristics, general health condition, potential side effects of treatment, and of course, patient's preference.

An extensive counseling and exchange by multidisciplinary teams is frequently employed at certain institutions. Certainly at Hopkins, we have multidisciplinary clinics that are focused on management of patients ranging from the active surveillance to all definitive treatments.

Between 20 and 42% of all screen detected cancers in the US are overtreated. The PSA detection has been found to be responsible for up to about seven years of lead time bias in terms of diagnosis of prostate cancer. What it means is that in a younger patient population, age 55, about 12 years of lead time bias is seen and in older patients up to six years.

Who Should Be Recommended for Active Surveillance?

It has been determined that men with very low risk prostate cancer and life expectancy less than 20 years, and those with so-called low risk cancer and life expectancy of less than 10.

You see this very low risk prostate cancer versus low risk cancer differentiation and different active surveillance programs actually utilize slightly different inclusion criteria. I will review those applied at Hopkins. And then Mazu Haer is going to review his Toronto program.

There is a range of patients who go into these management programs.

My colleague at Hopkins Pathologist Epstein introduced criteria to predict pathologically insignificant prostate cancer. We talk a lot about the significant prostate cancer. I want to now focus on what constitutes insignificant prostate cancer.

In terms of pathologic diagnosis, it will be clinical stage T1c, which is a non-palpable disease biopsy, Gleason score less or equal to six. In the US, Gleason score of six or grade group of one is considered the lowest diagnosed by pathologists as cancer, contrary to European studies where they may consider 3+2 categories of five.

Presence of disease in less than three, which would be two core biopsies. According to Epstein criteria, less or equal to 50% of cancer involvement of any core. We have two cores, and none more than 50% of cancer involvement. PSA density is being used by Epstein criteria of less than 0.15, where it really becomes very important how the volume of prostate is calculated.

We know there are some differences between the TRUS and MRI, and again, when PSA density is considered, it is important. The definition may change as we consider the volumetric assessment by MRI in the future.

There are some caveats, as we know from the RP data, about 8% of cancers that qualified as insignificant using this criteria that I mentioned, they were non-organ confined on postsurgical findings. There is some importance of finding the cancer at its extent. We already heard about the contributions of MRI and MRI target biopsies in this aspect.

Advantages and Disadvantages of Active Surveillance

What are the advantages of choosing active surveillance? Certainly, patients avoid side effects of definitive therapies and specifically when they are treated for small and indolent cancer, and they may retain the quality of life and normal activities. Of course, decreasing initial costs have been considered. However, active surveillance requires certain management steps that actually over time may be costly.

Disadvantages, of course, is a potential or threat of missed opportunity for cure. That has been looked at in different study settings. I am going to give you reference to one.

Also, there is a possibility of cancer progression or metastasis before treatment is being instituted while patient is being managed in active surveillance. There is possible need to use more complex treatment, multiple modalities with increasing side effects of these modalities when larger and more aggressive cancers are being eventually detected, possibility that nerve sparing during subsequent prostatectomy will be more difficult because there will be preexisting multiple biopsies that may induce some scarring.

Chance of nerve sparing surgery and preservation of potency after surgery may be decreasing. Patient of course may be anxious that he lives with cancer that is being monitored rather than treated. The patient will need to go through active monitoring with frequent exams, periodic biopsies. We do not know everything about the long-term biology of the cancers, and specifically in the untreated scenario.

These are the concerns that patients may have.

Watchful Waiting vs. Active Surveillance

You probably heard the terms of watchful waiting. How does watchful waiting differ from the active surveillance with intent to treat?

When we monitor patients, as I mentioned, using different strategies to detect significant cancer that needs intervention, that is the goal of active surveillance. However, in watchful waiting, actually treatment is only used to palliate once the cancer is advanced. This is typically used in older patients with comorbidities.

Essentially, watchful waiting is doing nothing until palliation is needed. Active surveillance is monitoring until there is reclassification and interventional treatment needs to be instituted.

Risk Categories for Active Surveillance

I mentioned the very low risk prostate cancer and low risk prostate cancer, both categories of patients included in the active surveillance.

In the very low risk prostate cancer, and certain that applies to our criteria as Hopkins has, Epstein component have life expectancy less than 20 years. Cancer is not palpated, stage T1c, PSA density is less than 0.15, Gleason score is no more than six. There is no pattern four and no more than two cores are found with cancer. This repeats what I mentioned on Epstein criteria.

The low risk prostate cancer life expectancy less than 10, 15 years, cancer again is not felt or perhaps a small nodule is felt. In T2a disease, PSA is considered in this scenario, and it is less than 10. Gleason score is six or less.

Differences Between Caucasian and African-American Men

There is a difference between Caucasian and African-American men, and this is important to consider. As we know, in African American men, clinically significant cancer is found not infrequently in the anterior portion of the gland. Actually in this population, these men have anterior dominant tumors, and also they have more likely significant cancer. In radical prostatectomies, they have higher stage, grade and volume disease.

We know that African-American men enrolled at Hopkins have twice the chance of progression on the repeat biopsy than Caucasian males. We also know that African-American men with Gleason score six disease have lower PSA and PSA density than Caucasians.

It is important to consider the ethnicity in patients as we counsel them and perhaps use MRI in our strategic portfolio.

Insignificant Disease Prediction vs. RP Results

How does the insignificant disease prediction on pathology compare to results from the RP? We know that this insignificant small volume of cancer less than 0.5 cc organ confined Gleason six in Caucasian male based on biopsy, based on original Epstein criteria.

We know that there is going to be a number of patients who are going to have an increase volume and Gleason grade at RP, and we know that about 3.5% in the Hopkins cohort had more adverse findings at RP. That is something to consider.

Among cancers to be predicted insignificant in African American men, there is quite significant presence again of anterior dominant tumors, which is 44% versus only 29% in Caucasian males. Also, 44, 54% of cases when we use the original Epstein criteria. There are some modifiers to it may be misclassified. Inclusion criteria differs.

Just wanted to tell you again that the number of cores percentage of positive cores differ between different programs. I explained to you what the Hopkins criteria are.

Surveillance Schedules and Progression Criteria

When we look at the surveillance schedules between the active surveillance, they differ as well. At Hopkins, we follow patients every six months with PSA and digital rectal exam, and we do a repeat biopsy every 12 months.

In other programs such as European PRIAS program, patients are seen every three months for two years, and their biopsy schedule differs. The progression criteria, meaning patient findings, trigger intervention, also are slightly differently defined in different programs.

At Hopkins, certainly if there is increase in number of positive biopsies over two, we have increase in volume. If there is increase Gleason grade over total score over six, we have also increase in the aggressiveness of the tumor. Other programs use again slightly different criteria, which you can figure out based on this very nice reference.

Quality of Life Assessment

Now when it comes to assessment of quality of life in different strategies for management of patients with prostate cancer. When you look at the active surveillance versus radical prostatectomy, this is a nice study presented at AUA.

You will see that in that study at two years after diagnosis, two cohorts, one on active surveillance, one in the radical prostatectomy arm at two years after diagnosis, sexual function declined in both. However, larger declines were seen in radical prostatectomy, and that was statistically significant.

At two years, urinary function declined again in both groups. However, there was larger decline in the radical prostatectomy group. There was no difference in terms of the bowel function, physical health or mental health between the group.

Compliance with Active Surveillance

Non-compliance with active surveillance is being looked at. In this study from Switzerland, about 27% of patients did not show up for the recommended appointments. Also, up to 11% of patients were lost to follow up in this study. About 50% of patients stay in the active surveillance group. Of course, some are reclassified, but some are non-compliant with the regimen.

Does Active Surveillance Miss the Window for Cure?

A very important aspect of not missing window of cure was looked at in this study that asks a question, does active surveillance miss the window for cure? Where matched comparison of immediate versus delayed prostatectomy was considered.

In this large study of 634 men who were matched with a controlled group, what was found that delayed radical prostatectomy was associated with a greater risk of an increased Gleason score above seven at time of surgery. However, there were no significant differences in biochemical recurrence, secondary treatment, or prostate cancer death between the groups. Considering that this was done only at seven years.

These are the evidence components that we know from the current research.

Hopkins Active Surveillance Program

At Hopkins, we have 1,125 men enrolled with the very low risk prostate cancer. We monitor them, as I mentioned before. In our cohort, 80% of patients match the criteria for very low risk disease. Another 20 may be enrolled because they wish to do so.

When we perform intervention is when there is disease reclassification. PSA density is considered above 0.15, Gleason score seven and above, more than two positive cores and increasing volume of prostate cancer more than 50%.

As of now, we have 56% of patients being active in the program, 38 had biopsy reclassification, 2% have died of causes other than prostate cancer, and 10% have either been lost to follow up or withdrew from the program.

Role of MRI in Active Surveillance

We can see that MRI can play a very interesting and important role in this cohort. You already mentioned that the detection of disease in unusual location, apex anterior transition zone are important aspects. We also mentioned the role of diffusion and apparent diffusion coefficients in prediction of higher grade tumors. These are important concepts.

In current practice at Hopkins, we refer eligible men for a multiparametric MRI before enrollment of active surveillance, and now we scan patients who are already in active surveillance.

We have done some studies and we already mentioned our worries about false negative results of prostate MRI. In our experience, the negative predictive value of prostate MRI multiparametric in the cohort of active surveillance patients is above about 89% with specificity of detecting pathological index tumor of 98%.

We did find that there is a difference between the MR visible and MR invisible tumors in terms of the patient's progression rate. What we found that MR invisibility of tumor was associated with a lower risk of adverse biopsy on subsequent pathology. You can see the relative risk was 0.2.

We have to say that in our cohort, 8.3% of invisible tumors had adverse pathology comparing to up to 40.5% of men who had visible tumors. Certainly tumor visibility has significant implications.

Case Studies from Hopkins

I am just going to show you three cases from our cohort.

71-year-old man who was followed in active surveillance for over 10 years. He had multiple cores of Gleason six disease, and we did visualize his tumor on T2 diffusion positive perfusion. We did spectroscopy choline is elevated. Of course, all the findings of positivity size was 12.

We scored T2 three because it was not really well defined. We scored diffusion as four because it was less than 0.5, and we scored positive DCE for an overall score of four. Patient remains on the active surveillance because of his age comorbidities, and he is in a follow-up MRI cohort.

Second patient, this is 60-year-old man who was considering active surveillance, had Gleason score three plus three biopsy, MRI detected anterior large positive tumor, no question in this patient, because we found this tumor, which would be size 21 by 42, score of five.

You can see invasive large tumor anterior transition zone diffusion score is of five, no question. This is positive overall score five, he underwent MR fusion biopsy. This is yet another system. This is a GE Logiq that we use in radiology at Hopkins.

You can see how we display the ultrasound and the MRI, we fuse this together or do it side by side. This large tumor was biopsied and it came back three plus four, and patient underwent radical prostatectomy.

The last case that I want to show you is this case of 71-year-old male who presented with PSA of 4.5. It increased from three. He had the negative DRE on biopsy. He had one core at three plus three 5% in right base. He also had a high grade PIN and some atypia.

Essentially he was offered active surveillance and based on clinical exam and certainly all these parameters, he would fit the patient was very anxious. Actually this patient demanded to have an MRI. So we did an MRI, and you have coronal axial T2 diffusion and perfusion.

I am asking you a question and answer to yourself, is prostate cancer visible? That is number one question. Can we see cancer? Can we identify the index nodule? Where is the index nodule? Is this patient a candidate for active surveillance?

When we look at the results, we say, the prostate cancer is visible because there are multiple foci of focal T2 signal decrease, matching diffusion restriction, and also matching positive DCE. We have a dominant nodule on the left side, mid posterior 15 millimeters, so that would be five.

There is little bit of capsular bulge, so the size really helped us here. ADC is score four, DCE is positive overall score of four because we rely on the ADC. Yes, cancer is visible. There is a dominant nodule in the left mid peripheral zone, posterior lateral.

This patient is not a candidate for active surveillance according to our Hopkins criteria, because now we have at least three different foci of cancer and one we consider clinically significant. He had robotic radical prostatectomy, final pathology, four plus three adenocarcinoma in the location as we specified.

Here is the tumor. He had nonfocal extra prostatic extension. This case illustrates that even cancers located in the peripheral zone, and we already heard about it may be missed on just blind TRUS, and this may be an important issue with these patients.

The false negative findings on prostate MRI in our cohort, about 89% in this cohort, nice paper by Park in British Journal of Urology revealed false negative in about 14.3%. This is something we need to understand. MRI is not perfect, but MRI really contributes important findings that we need to take into consideration.

These are the arguments when we exchange with our urology colleagues that we can bring to the table and working as a team. I would emphasize teamwork is the key.

At Hopkins, we have had the long sort of hate and now love affair with the prostate MRI. As I say, it took me from the time I started working with Jeff Weiner on AURIN 6659 trial, it took me over 10 years, consistent, persistent, and insistent participation in multidisciplinary clinic and working hard as we can to actually have very good relationship and scan more and more patients to their benefit, as I believe.

So thank you very much. We will have Mazu Hader talk about his program and demonstrate some cases live for you.

Toronto Active Surveillance Program

I am going to start just by showing four cases, two on slides, and two with the viewing software. Then get into some didactic stuff around the surveillance program in Canada.

Case 1: Anterior Midline Tumor

First case, patient is on active surveillance. PSA was 12, had gone up to 12 on active surveillance, 59 years old. The systematic biopsy had shown Gleason score six involving 25% of a core.

What I want to show you here is a pattern you now see multiple times. It is a similar pattern to the high PSA repeat negative biopsy pattern. This is not unusual in patients similarly on active surveillance who have biopsy for cause. These are these anterior midline tumors.

This particular tumor is apical, and we are seeing certainly this is an area where it is very sometimes hard to see on gray scale ultrasound, a lot of shadowing behind the urethra. Even the biopsy approach to this, when you go through the urethra can be quite painful for the patient. Requires a little more sort of angulation to get to it.

A fairly typical lesion here. The ADC is quite low, as is typical. I am not going to go on about this, but has a lot of implications, a lot of description in the report for this in terms of the surgical planning approach, the fact that it is apical, the fact that it is near the sphincter.

I spend more time in the report describing the location, the extraprostatic disease potential, and the potential for positive margins if the patient goes to surgery with these ones.

Case 2: Prostatitis and Tumor Progression

Second case, prostatitis in this patient known, with a PSA of 12 on entrance, 59 years old, at least Gleason six 25% of core.

This gets to, I think one of the things that we need more data on, and we need to have some established guidelines on, is that the game in active surveillance is to define the clinically significant cancer and find it in a timely fashion. So you can go after it.

In 2011, this patient comes and has a very obvious tumor here at the right base. PSA was 9.3 in 2011. Here is the tumor. We did targeted cores. We did three targeted cores through this lesion, along with a systematic biopsy. This showed Gleason six cancer in more than 50%.

In the Toronto active surveillance criteria, Gleason six is not necessarily actionable cancer in a 74-year-old man. Even though it is high volume six, it was elected to follow this patient. But the question is, when is the next MR going to happen? What are the criteria to do the next MR.

This patient had known prostatitis, so they actually had prostatitis on the biopsy acute prostatitis, and their PSA continued to fluctuate until just recently, and this is in the last two months, the patient came back and the PSA had just taken a big jump to 20, and we repeated the MR.

What you can see now is the tumor is bigger. Here is just the DWI image, even at this is just a b1000, so not a high b and very restricted, and the tumor is larger. We are now going to repeat this.

Two questions immediately came to mind when I saw this case is one, should we be doing MRs more frequently in these patients with high volume six? I think that the consensus is growing, although there is not a lot of evidence yet that we should be doing these MRs probably no less frequently than every two years.

The second thing is, how many cores should we be taking through the index lesion? Because I am asking myself the question, did we miss a four component even though we did take multiple cores in this patient.

What we are doing now at our center routinely is that if it is a fused biopsy through the index lesion, we are taking four cores. If we see a tumor clearly on MR, it is a PI-RADS four or five, we are taking four cores through that in the hopes of not missing the Gleason four component.

If it is a cognitive fusion, we are taking six, again, I do not have evidence for those thresholds, but that is what we are doing right now, just to give us reassurance that we are not missing the confirmatory biopsy on what the main game in active surveillance is, which is identifying based on Gleason score a clinically significant cancer.

I thought that would be good to illustrate.

Case 3: Transition Zone Tumor

We are going to switch to the workstation now. This patient PSA rise to 11 nanograms, 68-year-old on active surveillance. I just want to show you a transition zone tumor here.

This is a b1000 image. This is an extrapolated b2000 image showing this anterior tumor. You can see that it is a little noisy, this image, but you can see here the loss of the charcoal smudge sign or loss of texture here.

What I want to show you is just the ADC. I will admit that I still think the Achilles heel in a lot of our interpretation strategies right now is the transition zone tumor and really identifying it with confidence cases like this that are fairly obvious, I think we can all get, but a lot of the other ones I think are challenging.

One of the things I am doing right now is this is how I evaluate quantitative ADC, I look at, I use a probe tool and move through it because it tends to be very heterogeneous here. If you look, the numbers are ranging between about 500 and a thousand within the tumor. I find that helpful to just move the probe tool through.

I use the same window level setting, which is about 1400 by 1400 to view all my cases from this one particular scanner. I do find it helpful if it is less than about 650 in the transition zone, that makes me a little more worried. I take a second look at the T2, so it comes together that way.

This is showing up on the b1000 very clearly. It is got all the features, PI-RADS five, transition zone tumor. Sure enough, this was a clinically significant cancer on biopsy.

The ADC in terms of an ROI was 506, clearly abnormal Gleason four plus three tumor, 80% on targeted core.

Case 4: Small Peripheral Zone Lesion

This second case, this is the second case I will show using the workstation is a 69-year-old PSA 7.95, only 5% Gleason six on biopsy. Just zoom up here.

I am just going to scroll through this case and say, do you see a tumor here? Let me just fix the window level setting there. I am going to start at the apex and work up slowly and go back down.

This is a tough case. I actually do not think I could have easily picked this up based on these images. The tumor is actually right here. Here it is on the b1000 image. This is the extrapolated image. Whenever you see a white band like this on the extrapolated image, it means that there was significant motion. This means that your ADC values might be off as well.

Here is the lesion, but the ADC is not going to be as low as we would like it to be, simply because the patient moved. The calculated values are not going to be great. I am just going to show you the curve on this. If I can get that up.

I am just going to move through the prostate and you will see the curve shapes as I move through and as we get to the lesion, I can get it there. There we go. You can see again, lots of wash in, a little bit of wash out, so you can confirm that something.

How did I detect this? I am going through a stack of cases, and what I want to show you here is I will just bring up the other hanging protocol. This one is how important it is to look at all three planes on T2.

We say that ADC is king in the peripheral zone, but the reality is that the marriage here is very much between the T2 and ADC together. The T2 is your anchor. It is what is going to help you identify the lesion, the morphology of the lesion.

Getting good quality, I like to get all three planes. You should be doing at least coronal and axial imaging. These active surveillance for patients are exactly the patients where these small lesions you can miss them. They are low risk to start with.

Here is the tumor. It is very clear, it is a nodule, and the amount of restriction it is exhibiting for its size is quite significant, and you will get a feel for that as that goes on. That is why this was a definitive call of PI-RADS four.

If we could go back to the slides. There it is. At biopsy, this patient went from Gleason six 5% to Gleason three plus four 70%. We did do our four fused cores through this. In fact, every single one of the four cores was positive. The tumor is more extensive than is indicated by the size on imaging. There was also perineural invasion on biopsy.

Just a couple of illustrative cases in terms of how I am doing interpretation, and some of the difficulties with small lesions.

Spectrum of Active Surveillance in Toronto

Where are we in the spectrum of active surveillance in Toronto? In Toronto, Lori Klotz is a urologist who is well known in this field, but sits, I would say on the conservative camp in terms of follow up.

He really believes that the four component is the one that makes the tumor significant. He shifted in the last few years. This is before MR got introduced at his center, to really saying four plus three. I think he might revisit that. I do not know.

Biopsy is also not done on entrance. Many other surveillance programs, they do biopsy. As soon as the patient is a candidate for active surveillance, they will do an extended or 12 core or more biopsy.

In Canada, what we do is that they have already had a biopsy done somewhere else. It may not be a full 12 core, but we wait a year onto active surveillance before we then do an extended biopsy. Now what we are doing is we are introducing MR into that paradigm within the first year to guide that second biopsy.

We are actually doing a randomized control trial in Ontario with MR in active surveillance. To show you what the implications are of a very conservative intervention approach with quite the high bar on clinically significant disease at four plus three, the implications are very much the first case I showed you, that even though the guy had high volume Gleason six tumor had a sizable lesion on MR, it would be considered reasonable to follow this patient. 'Cause there is no four components. I really want to be sure that there is no four components as part of this paradigm.

That is where we are in the spectrum.

Outcomes of the Conservative Approach

What has happened to these patients with that conservative approach? In fact, if you look and this is 15 year follow up, there is a prostate cancer specific mortality in this group of 5% at 15 years, which raises some questions.

There is a five year biochemical recurrence rate in patients who undergo treatment. That is, whether it is prostatectomy or radiation of almost 50%. If you look at when those recurrences are happening or when these patients have significant cancer discovered after treatment, it is actually in a very early time period.

This is not something that, where the majority of cases are late, many of them are early in this surveillance period, which again, begs the question, if we had an MR early and found the clinically significant cancer, would these patients have never been on active surveillance in the first place?

This is the data from Toronto. I think it is all being revisited. Lori, Dr. Klotz is very, very keen on bringing MR full force into active surveillance to change the paradigm.

Studies on MR in Toronto Active Surveillance

If we look at our data in Toronto, there was a couple of studies published, one in 2012, and this was a prospective review done by our urology group. What this basically shows is that if you have a positive MR, that is, would be a PI-RADS four or five, we were not using PI-RADS during that period. It was not really out yet for us to use, but that is where the reclassification occurs.

If the MR is negative, the percent reclassification is small. I think this is the consistent theme is one of the big impacts of MR is in the negative predictive value.

The question comes on the positive side, you are going to reclassify men. The incremental yield in our group is sitting at around 10% over a extended biopsy or 12 core or more biopsy after the first year. There is an incremental benefit, but certainly a big benefit in terms of the negative MR and being sure about low risk.

This is another cross-sectional study we did and published again in men on active surveillance with this paradigm, using Artemis biopsy system. What this essentially shows is that when you have got PI-RADS one or two, and I think the question was asked by someone today, is what threshold should you use? Should it be three, should it be four for doing the biopsy?

What we found in our population is certainly if it is one or two, whether it is clinically significant disease, at a threshold of 50% of a core positive, at least Gleason six or clinically significant disease, at a threshold of any Gleason four at all, that if you have got PI-RADS one or two, the chance of this, this was about 80 patients, was close to zero, it was very, very low.

Now the number is going to be higher than zero, but PI-RADS one or two in this context already low risk patients is very, very low. Three, the number is still low. But when you get into the four or five, not unexpectedly, the yields are very high for clinically significant disease, whatever threshold you pick.

This is just confirming that certainly one or two can be very, very comfortable at spreading out that follow up interval by maybe two years. But if it is three, you definitely want to be doing follow up, at least follow up MR in addition to the PSA, again, I am not talking evidence-based, this is just based on the data that we have out there, and what we are doing in Toronto.

You want probably want to do follow up MR at least within two years. If it is four or five, you really want to be making really, really sure that there is not significant disease there by having a robust sampling strategy.

That is it. I finished a little bit early.